def test_resolved_in_next_release(self):
self.login_as(user=self.user)
project = self.create_project()
project.flags.has_releases = True
project.save()
group = self.create_group(project=project)
Release.get_or_create(
version='abcd',
project=project,
)
url = '/api/0/issues/{}/'.format(group.id)
response = self.client.put(
url, data={
'status': 'resolvedInNextRelease',
}
)
assert response.status_code == 200, response.content
group = Group.objects.get(
id=group.id,
project=group.project.id,
)
assert group.status == GroupStatus.RESOLVED
assert GroupResolution.objects.filter(
group=group,
).exists()
def finish_release(self, version, **values):
if not Release.is_valid_version(version):
raise HookValidationError('Invalid release version: %s' % version)
values.setdefault('date_released', timezone.now())
try:
with transaction.atomic():
release = Release.objects.create(
version=version, organization_id=self.project.organization_id, **values
)
except IntegrityError:
release = Release.objects.get(
version=version,
organization_id=self.project.organization_id,
)
release.update(**values)
release.add_project(self.project)
Activity.objects.create(
type=Activity.RELEASE,
project=self.project,
ident=Activity.get_version_ident(version),
data={'version': version},
datetime=values['date_released'],
)
self.set_refs(release=release, **values)
def start_release(self, version, **values):
values.setdefault('date_started', timezone.now())
affected = Release.objects.filter(
version=version,
organization_id=self.project.organization_id,
projects=self.project,
).update(**values)
if not affected:
release = Release.objects.filter(
version=version,
organization_id=self.project.organization_id,
).first()
if release:
release.update(**values)
else:
lock_key = Release.get_lock_key(self.project.organization_id, version)
lock = locks.get(lock_key, duration=5)
with TimedRetryPolicy(10)(lock.acquire):
try:
release = Release.objects.get(
version=version,
organization_id=self.project.organization_id
)
except Release.DoesNotExist:
release = Release.objects.create(
version=version,
organization_id=self.project.organization_id,
**values
)
release.add_project(self.project)
def get_release(self, project, data):
if not data.get('release'):
return
return Release.get(
project=project,
version=data['release'],
)
def get_release(self, create=False):
"""Convenient helper to return the release for the current data
and optionally creates the release if it's missing. In case there
is no release info it will return `None`.
"""
release = self.data.get('release')
if not release:
return None
if not create:
return Release.get(project=self.project, version=self.data['release'])
timestamp = self.data.get('timestamp')
if timestamp is not None:
date = datetime.fromtimestamp(timestamp).replace(tzinfo=timezone.utc)
else:
date = None
return Release.get_or_create(
project=self.project,
version=self.data['release'],
date_added=date,
)
def preprocess_step(self, processing_task):
frames = self.get_valid_frames()
if not frames:
logger.debug('Event %r has no frames with enough context to '
'fetch remote source', self.data['event_id'])
return False
if self.data.get('release'):
self.release = Release.get(
project=self.project,
version=self.data['release'],
)
self.populate_source_cache(frames)
return True
def start_release(self, version, **values):
if not Release.is_valid_version(version):
raise HookValidationError('Invalid release version: %s' % version)
try:
with transaction.atomic():
release = Release.objects.create(
version=version, organization_id=self.project.organization_id, **values
)
except IntegrityError:
release = Release.objects.get(
version=version,
organization_id=self.project.organization_id,
)
release.update(**values)
release.add_project(self.project)
def make_release_generator():
id_sequence = itertools.count(1)
while True:
dt = to_datetime(
random.randint(
timestamp - (30 * 24 * 60 * 60),
timestamp,
), )
p = random.choice(projects)
yield Release(
id=next(id_sequence),
project=p,
organization_id=p.organization_id,
version=''.join(
[random.choice('0123456789abcdef') for _ in range(40)]),
date_added=dt,
)
def start_release(self, version, **values):
if not Release.is_valid_version(version):
raise HookValidationError('Invalid release version: %s' % version)
try:
with transaction.atomic():
release = Release.objects.create(
version=version, organization_id=self.project.organization_id, **values
)
except IntegrityError:
release = Release.objects.get(
version=version,
organization_id=self.project.organization_id,
)
release.update(**values)
release.add_project(self.project)
def get(self, request):
org = Organization(
id=1,
slug='organization',
name='My Company',
)
team = Team(
id=1,
slug='team',
name='My Team',
organization=org,
)
project = Project(
id=1,
organization=org,
team=team,
slug='project',
name='My Project',
)
release = Release(
project=project,
version=sha1(uuid4().bytes).hexdigest(),
)
release_link = absolute_uri('/{}/{}/releases/{}/'.format(
org.slug,
project.slug,
release.version,
))
project_link = absolute_uri('/{}/{}/'.format(
org.slug,
project.slug,
))
return MailPreview(
html_template='sentry/emails/activity/release.html',
text_template='sentry/emails/activity/release.txt',
context={
'release': release,
'project': project,
'release_link': release_link,
'project_link': project_link,
},
).render(request)
def get_tag_value_label(self, key, value):
label = value
if key == "sentry:user":
if value.startswith("id:"):
label = value[len("id:"):]
elif value.startswith("email:"):
label = value[len("email:"):]
elif value.startswith("username:"******"username:"******"ip:"):
label = value[len("ip:"):]
elif key == "sentry:release":
from sentry.models import Release
label = Release.get_display_version(value)
return label
def get_tag_value_label(self, key, value):
label = value
if key == 'sentry:user':
if value.startswith('id:'):
label = value[len('id:'):]
elif value.startswith('email:'):
label = value[len('email:'):]
elif value.startswith('username:'******'username:'******'ip:'):
label = value[len('ip:'):]
elif key == 'sentry:release':
from sentry.models import Release
label = Release.get_display_version(value)
return label
def get_tag_value_label(self, key, value):
label = value
if key == 'sentry:user':
if value.startswith('id:'):
label = value[len('id:'):]
elif value.startswith('email:'):
label = value[len('email:'):]
elif value.startswith('username:'******'username:'******'ip:'):
label = value[len('ip:'):]
elif key == 'sentry:release':
from sentry.models import Release
label = Release.get_display_version(value)
return label
def debounce_update_release_health_data(organization, project_ids):
"""This causes a flush of snuba health data to the postgres tables once
per minute for the given projects.
"""
# Figure out which projects need to get updates from the snuba.
should_update = {}
cache_keys = ["debounce-health:%d" % id for id in project_ids]
cache_data = cache.get_many(cache_keys)
for project_id, cache_key in izip(project_ids, cache_keys):
if cache_data.get(cache_key) is None:
should_update[project_id] = cache_key
if not should_update:
return
projects = {
p.id: p
for p in Project.objects.get_many_from_cache(should_update.keys())
}
# This gives us updates for all release-projects which have seen new
# health data over the last 24 hours. It will miss releases where the last
# date is <24h ago. We need to aggregate the data for the totals per release
# manually here now. This does not take environments into account.
for project_id, version in get_changed_project_release_model_adoptions(
should_update.keys()):
project = projects.get(project_id)
if project is None:
# should not happen
continue
# We might have never observed the release. This for instance can
# happen if the release only had health data so far. For these cases
# we want to create the release the first time we observed it on the
# health side.
release = Release.get_or_create(project=project, version=version)
# Make sure that the release knows about this project. Like we had before
# the project might not have been associated with this release yet.
release.add_project(project)
# Debounce updates for a minute
cache.set_many(
dict(izip(should_update.values(), [True] * len(should_update))), 60)
def test_transfer_to_team_releases(self):
from_org = self.create_organization()
from_team = self.create_team(organization=from_org)
to_org = self.create_organization()
to_team = self.create_team(organization=to_org)
project = self.create_project(teams=[from_team])
environment = Environment.get_or_create(project, 'production')
release = Release.get_or_create(project=project, version='1.0')
ReleaseProjectEnvironment.objects.create(
project=project,
release=release,
environment=environment,
)
assert ReleaseProjectEnvironment.objects.filter(
project=project,
release=release,
environment=environment,
).exists()
assert ReleaseProject.objects.filter(
project=project,
release=release,
).exists()
project.transfer_to(team=to_team)
project = Project.objects.get(id=project.id)
assert project.teams.count() == 1
assert project.teams.first() == to_team
assert project.organization_id == to_org.id
assert not ReleaseProjectEnvironment.objects.filter(
project=project,
release=release,
environment=environment,
).exists()
assert not ReleaseProject.objects.filter(
project=project,
release=release,
).exists()
def test_transfer_to_team_releases(self):
from_org = self.create_organization()
from_team = self.create_team(organization=from_org)
to_org = self.create_organization()
to_team = self.create_team(organization=to_org)
project = self.create_project(teams=[from_team])
environment = Environment.get_or_create(project, 'production')
release = Release.get_or_create(project=project, version='1.0')
ReleaseProjectEnvironment.objects.create(
project=project,
release=release,
environment=environment,
)
assert ReleaseProjectEnvironment.objects.filter(
project=project,
release=release,
environment=environment,
).exists()
assert ReleaseProject.objects.filter(
project=project,
release=release,
).exists()
project.transfer_to(team=to_team)
project = Project.objects.get(id=project.id)
assert project.teams.count() == 1
assert project.teams.first() == to_team
assert project.organization_id == to_org.id
assert not ReleaseProjectEnvironment.objects.filter(
project=project,
release=release,
environment=environment,
).exists()
assert not ReleaseProject.objects.filter(
project=project,
release=release,
).exists()
def set_commits(self, version, commit_list):
"""
Commits should be ordered oldest to newest.
Calling this method will remove all existing commit history.
"""
if not Release.is_valid_version(version):
raise HookValidationError("Invalid release version: %s" % version)
project = self.project
try:
with transaction.atomic():
release = Release.objects.create(
organization_id=project.organization_id, version=version)
except IntegrityError:
release = Release.objects.get(
organization_id=project.organization_id, version=version)
release.add_project(project)
release.set_commits(commit_list)
def set_commits(self, version, commit_list):
"""
Commits should be ordered oldest to newest.
Calling this method will remove all existing commit history.
"""
if not Release.is_valid_version(version):
raise HookValidationError('Invalid release version: %s' % version)
project = self.project
try:
with transaction.atomic():
release = Release.objects.create(
organization_id=project.organization_id, version=version
)
except IntegrityError:
release = Release.objects.get(organization_id=project.organization_id, version=version)
release.add_project(project)
release.set_commits(commit_list)
def get_attrs(self, item_list, user):
result = {}
for item in item_list:
label = item.value
if item.key == 'sentry:user':
if item.value.startswith('id:'):
label = item.value[len('id:'):]
elif item.value.startswith('email:'):
label = item.value[len('email:'):]
elif item.value.startswith('username:'******'username:'******'ip:'):
label = item.value[len('ip:'):]
elif item.key == 'sentry:release':
label = Release.get_display_version(item.value)
result[item] = {
'name': label,
}
return result
def finish_release(self, version, **values):
values.setdefault('date_released', timezone.now())
affected = Release.objects.filter(
version=version,
organization_id=self.project.organization_id,
projects=self.project,
).update(**values)
if not affected:
release = Release.objects.filter(
version=version,
organization_id=self.project.organization_id,
).first()
if release:
release.update(**values)
else:
lock_key = Release.get_lock_key(self.project.organization_id,
version)
lock = locks.get(lock_key, duration=5)
with TimedRetryPolicy(10)(lock.acquire):
try:
release = Release.objects.get(
version=version,
organization_id=self.project.organization_id,
)
except Release.DoesNotExist:
release = Release.objects.create(
version=version,
organization_id=self.project.organization_id,
**values)
release.add_project(self.project)
activity = Activity.objects.create(
type=Activity.RELEASE,
project=self.project,
ident=version,
data={'version': version},
datetime=values['date_released'],
)
activity.send_notification()
def finish_release(self, version, **values):
values.setdefault('date_released', timezone.now())
affected = Release.objects.filter(
version=version,
organization_id=self.project.organization_id,
projects=self.project,
).update(**values)
if not affected:
release = Release.objects.filter(
version=version,
organization_id=self.project.organization_id,
).first()
if release:
release.update(**values)
else:
lock_key = Release.get_lock_key(self.project.organization_id, version)
lock = locks.get(lock_key, duration=5)
with TimedRetryPolicy(10)(lock.acquire):
try:
release = Release.objects.get(
version=version,
organization_id=self.project.organization_id,
)
except Release.DoesNotExist:
release = Release.objects.create(
version=version,
organization_id=self.project.organization_id,
**values
)
release.add_project(self.project)
activity = Activity.objects.create(
type=Activity.RELEASE,
project=self.project,
ident=version,
data={'version': version},
datetime=values['date_released'],
)
activity.send_notification()
def test_commits_lock_conflict(self):
user = self.create_user(is_staff=False, is_superuser=False)
org = self.create_organization()
org.flags.allow_joinleave = False
org.save()
team = self.create_team(organization=org)
project = self.create_project(name="foo",
organization=org,
teams=[team])
self.create_member(teams=[team], user=user, organization=org)
self.login_as(user=user)
release = self.create_release(project, self.user, version="1.2.1")
release.add_project(project)
# Simulate a concurrent request by using an existing release
# that has its commit lock taken out.
lock = locks.get(Release.get_lock_key(org.id, release.id), duration=10)
lock.acquire()
url = reverse(
"sentry-api-0-organization-release-details",
kwargs={
"organization_slug": org.slug,
"version": release.version
},
)
response = self.client.put(
url, data={"commits": [{
"id": "a" * 40
}, {
"id": "b" * 40
}]})
assert response.status_code == 409, (response.status_code,
response.content)
assert "Release commits" in response.data["detail"]
def test_simple(self):
date = datetime.datetime.utcnow()
org = self.create_organization()
project = self.create_project(organization=org, name='foo')
# this shouldn't be included
release1 = Release.objects.create(
organization=org,
version='a' * 40,
date_released=date - datetime.timedelta(days=2),
)
release1.add_project(project)
release2 = Release.objects.create(
organization=org,
version='b' * 40,
date_released=date - datetime.timedelta(days=1),
)
release2.add_project(project)
release3 = Release.objects.create(
organization=org,
version='c' * 40,
date_released=date,
)
release3.add_project(project)
releases = list(Release.get_closest_releases(project,
release2.version))
assert len(releases) == 2
assert releases[0] == release2
assert releases[1] == release3
def _get_or_create_release_many(jobs, projects):
jobs_with_releases = {}
release_date_added = {}
for job in jobs:
if not job["release"]:
continue
release_key = (job["project_id"], job["release"])
jobs_with_releases.setdefault(release_key, []).append(job)
new_datetime = job["event"].datetime
old_datetime = release_date_added.get(release_key)
if old_datetime is None or new_datetime > old_datetime:
release_date_added[release_key] = new_datetime
for (project_id,
version), jobs_to_update in six.iteritems(jobs_with_releases):
release = Release.get_or_create(
project=projects[project_id],
version=version,
date_added=release_date_added[(project_id, version)],
)
for job in jobs_to_update:
# dont allow a conflicting 'release' tag
data = job["data"]
pop_tag(data, "release")
set_tag(data, "sentry:release", release.version)
job["release"] = release
if job["dist"]:
job["dist"] = job["release"].add_dist(job["dist"],
job["event"].datetime)
# dont allow a conflicting 'dist' tag
pop_tag(job["data"], "dist")
set_tag(job["data"], "sentry:dist", job["dist"].name)
def test_simple(self):
date = datetime.datetime.utcnow()
org = self.create_organization()
project = self.create_project(organization=org, name='foo')
# this shouldn't be included
release1 = Release.objects.create(
organization=org,
version='a' * 40,
date_released=date - datetime.timedelta(days=2),
)
release1.add_project(project)
release2 = Release.objects.create(
organization=org,
version='b' * 40,
date_released=date - datetime.timedelta(days=1),
)
release2.add_project(project)
release3 = Release.objects.create(
organization=org,
version='c' * 40,
date_released=date,
)
release3.add_project(project)
releases = list(Release.get_closest_releases(project, release2.version))
assert len(releases) == 2
assert releases[0] == release2
assert releases[1] == release3
def get_latest_release(
projects: Sequence[Project | int],
environments: Optional[Sequence[Environment]],
organization_id: Optional[int] = None,
) -> Sequence[str]:
if organization_id is None:
project = projects[0]
if hasattr(project, "organization_id"):
organization_id = project.organization_id
else:
return []
# Convert projects to ids so that we can work with them more easily
project_ids = [getattr(project, "id", project) for project in projects]
semver_project_ids = []
date_project_ids = []
for project_id in project_ids:
if follows_semver_versioning_scheme(organization_id, project_id):
semver_project_ids.append(project_id)
else:
date_project_ids.append(project_id)
versions = set()
versions.update(
_run_latest_release_query(LatestReleaseOrders.SEMVER,
semver_project_ids, environments,
organization_id))
versions.update(
_run_latest_release_query(LatestReleaseOrders.DATE, date_project_ids,
environments, organization_id))
if not versions:
raise Release.DoesNotExist()
return list(sorted(versions))
def ensure_release_exists(instance, created, **kwargs):
if instance.key != 'sentry:release':
return
if instance.data and instance.data.get('release_id'):
return
affected = Release.objects.filter(
organization_id=instance.project.organization_id,
version=instance.value,
projects=instance.project
).update(date_added=instance.first_seen)
if not affected:
release = Release.objects.filter(
organization_id=instance.project.organization_id,
version=instance.value
).first()
if release:
release.update(date_added=instance.first_seen)
else:
lock_key = Release.get_lock_key(instance.project.organization_id, instance.value)
lock = locks.get(lock_key, duration=5)
with TimedRetryPolicy(10)(lock.acquire):
try:
release = Release.objects.get(
organization_id=instance.project.organization_id,
version=instance.value,
)
except Release.DoesNotExist:
release = Release.objects.create(
organization_id=instance.project.organization_id,
version=instance.value,
date_added=instance.first_seen,
)
instance.update(data={'release_id': release.id})
release.add_project(instance.project)
def setUp(self):
self.project = self.create_project()
self.release = Release.get_or_create(self.project, '1.0')
self.environment1 = Environment.get_or_create(self.project, 'prod')
self.environment2 = Environment.get_or_create(self.project, 'staging')
self.timestamp = 1403007314
def save(self, project, raw=False):
from sentry.tasks.post_process import index_event_tags
project = Project.objects.get_from_cache(id=project)
data = self.data.copy()
# First we pull out our top-level (non-data attr) kwargs
event_id = data.pop('event_id')
level = data.pop('level')
culprit = data.pop('culprit', None)
logger_name = data.pop('logger', None)
server_name = data.pop('server_name', None)
site = data.pop('site', None)
checksum = data.pop('checksum', None)
fingerprint = data.pop('fingerprint', None)
platform = data.pop('platform', None)
release = data.pop('release', None)
environment = data.pop('environment', None)
# unused
time_spent = data.pop('time_spent', None)
message = data.pop('message', '')
if not culprit:
# if we generate an implicit culprit, lets not call it a
# transaction
transaction_name = None
culprit = generate_culprit(data, platform=platform)
else:
transaction_name = culprit
date = datetime.fromtimestamp(data.pop('timestamp'))
date = date.replace(tzinfo=timezone.utc)
kwargs = {
'platform': platform,
}
event = Event(
project_id=project.id,
event_id=event_id,
data=data,
time_spent=time_spent,
datetime=date,
**kwargs
)
# convert this to a dict to ensure we're only storing one value per key
# as most parts of Sentry dont currently play well with multiple values
tags = dict(data.get('tags') or [])
tags['level'] = LOG_LEVELS[level]
if logger_name:
tags['logger'] = logger_name
if server_name:
tags['server_name'] = server_name
if site:
tags['site'] = site
if environment:
tags['environment'] = environment
if transaction_name:
tags['transaction'] = transaction_name
if release:
# dont allow a conflicting 'release' tag
if 'release' in tags:
del tags['release']
tags['sentry:release'] = release
event_user = self._get_event_user(project, data)
if event_user:
# dont allow a conflicting 'user' tag
if 'user' in tags:
del tags['user']
tags['sentry:user'] = event_user.tag_value
for plugin in plugins.for_project(project, version=None):
added_tags = safe_execute(plugin.get_tags, event,
_with_transaction=False)
if added_tags:
# plugins should not override user provided tags
for key, value in added_tags:
tags.setdefault(key, value)
# tags are stored as a tuple
tags = tags.items()
# XXX(dcramer): we're relying on mutation of the data object to ensure
# this propagates into Event
data['tags'] = tags
data['fingerprint'] = fingerprint or ['{{ default }}']
for path, iface in six.iteritems(event.interfaces):
data['tags'].extend(iface.iter_tags())
# Get rid of ephemeral interface data
if iface.ephemeral:
data.pop(iface.get_path(), None)
# prioritize fingerprint over checksum as its likely the client defaulted
# a checksum whereas the fingerprint was explicit
if fingerprint:
hashes = [
md5_from_hash(h)
for h in get_hashes_from_fingerprint(event, fingerprint)
]
elif checksum:
hashes = [checksum]
data['checksum'] = checksum
else:
hashes = [
md5_from_hash(h)
for h in get_hashes_for_event(event)
]
# TODO(dcramer): temp workaround for complexity
data['message'] = message
event_type = eventtypes.get(data.get('type', 'default'))(data)
event_metadata = event_type.get_metadata()
# TODO(dcramer): temp workaround for complexity
del data['message']
data['type'] = event_type.key
data['metadata'] = event_metadata
# index components into ``Event.message``
# See GH-3248
if event_type.key != 'default':
if 'sentry.interfaces.Message' in data and \
data['sentry.interfaces.Message']['message'] != message:
message = u'{} {}'.format(
message,
data['sentry.interfaces.Message']['message'],
)
if not message:
message = ''
elif not isinstance(message, six.string_types):
message = force_text(message)
for value in six.itervalues(event_metadata):
value_u = force_text(value, errors='replace')
if value_u not in message:
message = u'{} {}'.format(message, value_u)
if culprit and culprit not in message:
culprit_u = force_text(culprit, errors='replace')
message = u'{} {}'.format(message, culprit_u)
message = trim(message.strip(), settings.SENTRY_MAX_MESSAGE_LENGTH)
event.message = message
kwargs['message'] = message
group_kwargs = kwargs.copy()
group_kwargs.update({
'culprit': culprit,
'logger': logger_name,
'level': level,
'last_seen': date,
'first_seen': date,
'active_at': date,
'data': {
'last_received': event.data.get('received') or float(event.datetime.strftime('%s')),
'type': event_type.key,
# we cache the events metadata on the group to ensure its
# accessible in the stream
'metadata': event_metadata,
},
})
if release:
release = Release.get_or_create(
project=project,
version=release,
date_added=date,
)
group_kwargs['first_release'] = release
group, is_new, is_regression, is_sample = self._save_aggregate(
event=event,
hashes=hashes,
release=release,
**group_kwargs
)
event.group = group
# store a reference to the group id to guarantee validation of isolation
event.data.bind_ref(event)
try:
with transaction.atomic(using=router.db_for_write(EventMapping)):
EventMapping.objects.create(
project=project, group=group, event_id=event_id)
except IntegrityError:
self.logger.info('duplicate.found', extra={'event_id': event.id}, exc_info=True)
return event
environment = Environment.get_or_create(
project=project,
name=environment,
)
if release:
ReleaseEnvironment.get_or_create(
project=project,
release=release,
environment=environment,
datetime=date,
)
grouprelease = GroupRelease.get_or_create(
group=group,
release=release,
environment=environment,
datetime=date,
)
counters = [
(tsdb.models.group, group.id),
(tsdb.models.project, project.id),
]
if release:
counters.append((tsdb.models.release, release.id))
tsdb.incr_multi(counters, timestamp=event.datetime)
frequencies = [
# (tsdb.models.frequent_projects_by_organization, {
# project.organization_id: {
# project.id: 1,
# },
# }),
# (tsdb.models.frequent_issues_by_project, {
# project.id: {
# group.id: 1,
# },
# })
(tsdb.models.frequent_environments_by_group, {
group.id: {
environment.id: 1,
},
})
]
if release:
frequencies.append(
(tsdb.models.frequent_releases_by_group, {
group.id: {
grouprelease.id: 1,
},
})
)
tsdb.record_frequency_multi(frequencies, timestamp=event.datetime)
UserReport.objects.filter(
project=project, event_id=event_id,
).update(group=group)
# save the event unless its been sampled
if not is_sample:
try:
with transaction.atomic(using=router.db_for_write(Event)):
event.save()
except IntegrityError:
self.logger.info('duplicate.found', extra={'event_id': event.id}, exc_info=True)
return event
index_event_tags.delay(
project_id=project.id,
group_id=group.id,
event_id=event.id,
tags=tags,
)
if event_user:
tsdb.record_multi((
(tsdb.models.users_affected_by_group, group.id, (event_user.tag_value,)),
(tsdb.models.users_affected_by_project, project.id, (event_user.tag_value,)),
), timestamp=event.datetime)
if is_new and release:
buffer.incr(Release, {'new_groups': 1}, {
'id': release.id,
})
safe_execute(Group.objects.add_tags, group, tags,
_with_transaction=False)
if not raw:
if not project.first_event:
project.update(first_event=date)
first_event_received.send(project=project, group=group, sender=Project)
post_process_group.delay(
group=group,
event=event,
is_new=is_new,
is_sample=is_sample,
is_regression=is_regression,
)
else:
self.logger.info('post_process.skip.raw_event', extra={'event_id': event.id})
# TODO: move this to the queue
if is_regression and not raw:
regression_signal.send_robust(sender=Group, instance=group)
return event
def save(self,
project_id,
raw=False,
assume_normalized=False,
cache_key=None):
"""
We re-insert events with duplicate IDs into Snuba, which is responsible
for deduplicating events. Since deduplication in Snuba is on the primary
key (based on event ID, project ID and day), events with same IDs are only
deduplicated if their timestamps fall on the same day. The latest event
always wins and overwrites the value of events received earlier in that day.
Since we increment counters and frequencies here before events get inserted
to eventstream these numbers may be larger than the total number of
events if we receive duplicate event IDs that fall on the same day
(that do not hit cache first).
"""
# Normalize if needed
if not self._normalized:
if not assume_normalized:
self.normalize()
self._normalized = True
data = self._data
project = Project.objects.get_from_cache(id=project_id)
project._organization_cache = Organization.objects.get_from_cache(
id=project.organization_id)
# Pull out the culprit
culprit = self.get_culprit()
# Pull the toplevel data we're interested in
level = data.get("level")
# TODO(mitsuhiko): this code path should be gone by July 2018.
# This is going to be fine because no code actually still depends
# on integers here. When we need an integer it will be converted
# into one later. Old workers used to send integers here.
if level is not None and isinstance(level, six.integer_types):
level = LOG_LEVELS[level]
transaction_name = data.get("transaction")
logger_name = data.get("logger")
release = data.get("release")
dist = data.get("dist")
environment = data.get("environment")
recorded_timestamp = data.get("timestamp")
# We need to swap out the data with the one internal to the newly
# created event object
event = self._get_event_instance(project_id=project_id)
self._data = data = event.data.data
event._project_cache = project
date = event.datetime
platform = event.platform
event_id = event.event_id
if transaction_name:
transaction_name = force_text(transaction_name)
# Right now the event type is the signal to skip the group. This
# is going to change a lot.
if event.get_event_type() == "transaction":
issueless_event = True
else:
issueless_event = False
# Some of the data that are toplevel attributes are duplicated
# into tags (logger, level, environment, transaction). These are
# different from legacy attributes which are normalized into tags
# ahead of time (site, server_name).
setdefault_path(data, "tags", value=[])
set_tag(data, "level", level)
if logger_name:
set_tag(data, "logger", logger_name)
if environment:
set_tag(data, "environment", environment)
if transaction_name:
set_tag(data, "transaction", transaction_name)
if release:
# dont allow a conflicting 'release' tag
pop_tag(data, "release")
release = Release.get_or_create(project=project,
version=release,
date_added=date)
set_tag(data, "sentry:release", release.version)
if dist and release:
dist = release.add_dist(dist, date)
# dont allow a conflicting 'dist' tag
pop_tag(data, "dist")
set_tag(data, "sentry:dist", dist.name)
else:
dist = None
event_user = self._get_event_user(project, data)
if event_user:
# dont allow a conflicting 'user' tag
pop_tag(data, "user")
set_tag(data, "sentry:user", event_user.tag_value)
# At this point we want to normalize the in_app values in case the
# clients did not set this appropriately so far.
grouping_config = load_grouping_config(
get_grouping_config_dict_for_event_data(data, project))
normalize_stacktraces_for_grouping(data, grouping_config)
for plugin in plugins.for_project(project, version=None):
added_tags = safe_execute(plugin.get_tags,
event,
_with_transaction=False)
if added_tags:
# plugins should not override user provided tags
for key, value in added_tags:
if get_tag(data, key) is None:
set_tag(data, key, value)
for path, iface in six.iteritems(event.interfaces):
for k, v in iface.iter_tags():
set_tag(data, k, v)
# Get rid of ephemeral interface data
if iface.ephemeral:
data.pop(iface.path, None)
# The active grouping config was put into the event in the
# normalize step before. We now also make sure that the
# fingerprint was set to `'{{ default }}' just in case someone
# removed it from the payload. The call to get_hashes will then
# look at `grouping_config` to pick the right parameters.
data["fingerprint"] = data.get("fingerprint") or ["{{ default }}"]
apply_server_fingerprinting(
data, get_fingerprinting_config_for_project(project))
# Here we try to use the grouping config that was requested in the
# event. If that config has since been deleted (because it was an
# experimental grouping config) we fall back to the default.
try:
hashes = event.get_hashes()
except GroupingConfigNotFound:
data["grouping_config"] = get_grouping_config_dict_for_project(
project)
hashes = event.get_hashes()
data["hashes"] = hashes
# we want to freeze not just the metadata and type in but also the
# derived attributes. The reason for this is that we push this
# data into kafka for snuba processing and our postprocessing
# picks up the data right from the snuba topic. For most usage
# however the data is dynamically overridden by Event.title and
# Event.location (See Event.as_dict)
materialized_metadata = self.materialize_metadata()
data.update(materialized_metadata)
data["culprit"] = culprit
received_timestamp = event.data.get("received") or float(
event.datetime.strftime("%s"))
if not issueless_event:
# The group gets the same metadata as the event when it's flushed but
# additionally the `last_received` key is set. This key is used by
# _save_aggregate.
group_metadata = dict(materialized_metadata)
group_metadata["last_received"] = received_timestamp
kwargs = {
"platform": platform,
"message": event.search_message,
"culprit": culprit,
"logger": logger_name,
"level": LOG_LEVELS_MAP.get(level),
"last_seen": date,
"first_seen": date,
"active_at": date,
"data": group_metadata,
}
if release:
kwargs["first_release"] = release
try:
group, is_new, is_regression = self._save_aggregate(
event=event, hashes=hashes, release=release, **kwargs)
except HashDiscarded:
event_discarded.send_robust(project=project,
sender=EventManager)
metrics.incr(
"events.discarded",
skip_internal=True,
tags={
"organization_id": project.organization_id,
"platform": platform
},
)
raise
else:
event_saved.send_robust(project=project,
event_size=event.size,
sender=EventManager)
event.group = group
else:
group = None
is_new = False
is_regression = False
event_saved.send_robust(project=project,
event_size=event.size,
sender=EventManager)
# store a reference to the group id to guarantee validation of isolation
event.data.bind_ref(event)
environment = Environment.get_or_create(project=project,
name=environment)
if group:
group_environment, is_new_group_environment = GroupEnvironment.get_or_create(
group_id=group.id,
environment_id=environment.id,
defaults={"first_release": release if release else None},
)
else:
is_new_group_environment = False
if release:
ReleaseEnvironment.get_or_create(project=project,
release=release,
environment=environment,
datetime=date)
ReleaseProjectEnvironment.get_or_create(project=project,
release=release,
environment=environment,
datetime=date)
if group:
grouprelease = GroupRelease.get_or_create(
group=group,
release=release,
environment=environment,
datetime=date)
counters = [(tsdb.models.project, project.id)]
if group:
counters.append((tsdb.models.group, group.id))
if release:
counters.append((tsdb.models.release, release.id))
tsdb.incr_multi(counters,
timestamp=event.datetime,
environment_id=environment.id)
frequencies = []
if group:
frequencies.append((tsdb.models.frequent_environments_by_group, {
group.id: {
environment.id: 1
}
}))
if release:
frequencies.append((tsdb.models.frequent_releases_by_group, {
group.id: {
grouprelease.id: 1
}
}))
if frequencies:
tsdb.record_frequency_multi(frequencies, timestamp=event.datetime)
if group:
UserReport.objects.filter(project=project,
event_id=event_id).update(
group=group, environment=environment)
# Enusre the _metrics key exists. This is usually created during
# and prefilled with ingestion sizes.
event_metrics = event.data.get("_metrics") or {}
event.data["_metrics"] = event_metrics
# Capture the actual size that goes into node store.
event_metrics["bytes.stored.event"] = len(
json.dumps(dict(event.data.items())))
# Load attachments first, but persist them at the very last after
# posting to eventstream to make sure all counters and eventstream are
# incremented for sure.
attachments = self.get_attachments(cache_key, event)
for attachment in attachments:
key = "bytes.stored.%s" % (attachment.type, )
event_metrics[key] = (event_metrics.get(key) or 0) + len(
attachment.data)
# Write the event to Nodestore
event.data.save()
if event_user:
counters = [(tsdb.models.users_affected_by_project, project.id,
(event_user.tag_value, ))]
if group:
counters.append((tsdb.models.users_affected_by_group, group.id,
(event_user.tag_value, )))
tsdb.record_multi(counters,
timestamp=event.datetime,
environment_id=environment.id)
if release:
if is_new:
buffer.incr(
ReleaseProject,
{"new_groups": 1},
{
"release_id": release.id,
"project_id": project.id
},
)
if is_new_group_environment:
buffer.incr(
ReleaseProjectEnvironment,
{"new_issues_count": 1},
{
"project_id": project.id,
"release_id": release.id,
"environment_id": environment.id,
},
)
if not raw:
if not project.first_event:
project.update(first_event=date)
first_event_received.send_robust(project=project,
event=event,
sender=Project)
eventstream.insert(
group=group,
event=event,
is_new=is_new,
is_regression=is_regression,
is_new_group_environment=is_new_group_environment,
primary_hash=hashes[0],
# We are choosing to skip consuming the event back
# in the eventstream if it's flagged as raw.
# This means that we want to publish the event
# through the event stream, but we don't care
# about post processing and handling the commit.
skip_consume=raw,
)
# Do this last to ensure signals get emitted even if connection to the
# file store breaks temporarily.
self.save_attachments(attachments, event)
metric_tags = {"from_relay": "_relay_processed" in self._data}
metrics.timing("events.latency",
received_timestamp - recorded_timestamp,
tags=metric_tags)
metrics.timing("events.size.data.post_save",
event.size,
tags=metric_tags)
metrics.incr(
"events.post_save.normalize.errors",
amount=len(self._data.get("errors") or ()),
tags=metric_tags,
)
return event
def save(self, project, raw=False):
from sentry.tasks.post_process import index_event_tags
data = self.data
project = Project.objects.get_from_cache(id=project)
# Check to make sure we're not about to do a bunch of work that's
# already been done if we've processed an event with this ID. (This
# isn't a perfect solution -- this doesn't handle ``EventMapping`` and
# there's a race condition between here and when the event is actually
# saved, but it's an improvement. See GH-7677.)
try:
event = Event.objects.get(
project_id=project.id,
event_id=data['event_id'],
)
except Event.DoesNotExist:
pass
else:
self.logger.info(
'duplicate.found',
exc_info=True,
extra={
'event_uuid': data['event_id'],
'project_id': project.id,
'model': Event.__name__,
}
)
return event
# First we pull out our top-level (non-data attr) kwargs
event_id = data.pop('event_id')
level = data.pop('level')
transaction_name = data.pop('transaction', None)
culprit = data.pop('culprit', None)
logger_name = data.pop('logger', None)
server_name = data.pop('server_name', None)
site = data.pop('site', None)
checksum = data.pop('checksum', None)
fingerprint = data.pop('fingerprint', None)
platform = data.pop('platform', None)
release = data.pop('release', None)
dist = data.pop('dist', None)
environment = data.pop('environment', None)
# unused
time_spent = data.pop('time_spent', None)
message = data.pop('message', '')
if not culprit:
if transaction_name:
culprit = transaction_name
else:
culprit = generate_culprit(data, platform=platform)
culprit = force_text(culprit)
if transaction_name:
transaction_name = force_text(transaction_name)
recorded_timestamp = data.pop('timestamp')
date = datetime.fromtimestamp(recorded_timestamp)
date = date.replace(tzinfo=timezone.utc)
kwargs = {
'platform': platform,
}
event = Event(
project_id=project.id,
event_id=event_id,
data=data,
time_spent=time_spent,
datetime=date,
**kwargs
)
event._project_cache = project
data = event.data.data
# convert this to a dict to ensure we're only storing one value per key
# as most parts of Sentry dont currently play well with multiple values
tags = dict(data.get('tags') or [])
tags['level'] = LOG_LEVELS[level]
if logger_name:
tags['logger'] = logger_name
if server_name:
tags['server_name'] = server_name
if site:
tags['site'] = site
if environment:
tags['environment'] = environment
if transaction_name:
tags['transaction'] = transaction_name
if release:
# dont allow a conflicting 'release' tag
if 'release' in tags:
del tags['release']
release = Release.get_or_create(
project=project,
version=release,
date_added=date,
)
tags['sentry:release'] = release.version
if dist and release:
dist = release.add_dist(dist, date)
tags['sentry:dist'] = dist.name
else:
dist = None
event_user = self._get_event_user(project, data)
if event_user:
# dont allow a conflicting 'user' tag
if 'user' in tags:
del tags['user']
tags['sentry:user'] = event_user.tag_value
# At this point we want to normalize the in_app values in case the
# clients did not set this appropriately so far.
normalize_in_app(data)
for plugin in plugins.for_project(project, version=None):
added_tags = safe_execute(plugin.get_tags, event, _with_transaction=False)
if added_tags:
# plugins should not override user provided tags
for key, value in added_tags:
tags.setdefault(key, value)
for path, iface in six.iteritems(event.interfaces):
for k, v in iface.iter_tags():
tags[k] = v
# Get rid of ephemeral interface data
if iface.ephemeral:
data.pop(iface.get_path(), None)
# tags are stored as a tuple
tags = tags.items()
data['tags'] = tags
data['fingerprint'] = fingerprint or ['{{ default }}']
# prioritize fingerprint over checksum as its likely the client defaulted
# a checksum whereas the fingerprint was explicit
if fingerprint:
hashes = [md5_from_hash(h) for h in get_hashes_from_fingerprint(event, fingerprint)]
elif checksum:
if HASH_RE.match(checksum):
hashes = [checksum]
else:
hashes = [md5_from_hash([checksum]), checksum]
data['checksum'] = checksum
else:
hashes = [md5_from_hash(h) for h in get_hashes_for_event(event)]
# TODO(dcramer): temp workaround for complexity
data['message'] = message
event_type = eventtypes.get(data.get('type', 'default'))(data)
event_metadata = event_type.get_metadata()
# TODO(dcramer): temp workaround for complexity
del data['message']
data['type'] = event_type.key
data['metadata'] = event_metadata
# index components into ``Event.message``
# See GH-3248
if event_type.key != 'default':
if 'sentry.interfaces.Message' in data and \
data['sentry.interfaces.Message']['message'] != message:
message = u'{} {}'.format(
message,
data['sentry.interfaces.Message']['message'],
)
if not message:
message = ''
elif not isinstance(message, six.string_types):
message = force_text(message)
for value in six.itervalues(event_metadata):
value_u = force_text(value, errors='replace')
if value_u not in message:
message = u'{} {}'.format(message, value_u)
if culprit and culprit not in message:
culprit_u = force_text(culprit, errors='replace')
message = u'{} {}'.format(message, culprit_u)
message = trim(message.strip(), settings.SENTRY_MAX_MESSAGE_LENGTH)
event.message = message
kwargs['message'] = message
received_timestamp = event.data.get('received') or float(event.datetime.strftime('%s'))
group_kwargs = kwargs.copy()
group_kwargs.update(
{
'culprit': culprit,
'logger': logger_name,
'level': level,
'last_seen': date,
'first_seen': date,
'active_at': date,
'data': {
'last_received': received_timestamp,
'type':
event_type.key,
# we cache the events metadata on the group to ensure its
# accessible in the stream
'metadata':
event_metadata,
},
}
)
if release:
group_kwargs['first_release'] = release
try:
group, is_new, is_regression, is_sample = self._save_aggregate(
event=event, hashes=hashes, release=release, **group_kwargs
)
except HashDiscarded:
event_discarded.send_robust(
project=project,
sender=EventManager,
)
metrics.incr(
'events.discarded',
skip_internal=True,
tags={
'organization_id': project.organization_id,
'platform': platform,
},
)
raise
else:
event_saved.send_robust(
project=project,
event_size=event.size,
sender=EventManager,
)
event.group = group
# store a reference to the group id to guarantee validation of isolation
event.data.bind_ref(event)
# When an event was sampled, the canonical source of truth
# is the EventMapping table since we aren't going to be writing out an actual
# Event row. Otherwise, if the Event isn't being sampled, we can safely
# rely on the Event table itself as the source of truth and ignore
# EventMapping since it's redundant information.
if is_sample:
try:
with transaction.atomic(using=router.db_for_write(EventMapping)):
EventMapping.objects.create(project=project, group=group, event_id=event_id)
except IntegrityError:
self.logger.info(
'duplicate.found',
exc_info=True,
extra={
'event_uuid': event_id,
'project_id': project.id,
'group_id': group.id,
'model': EventMapping.__name__,
}
)
return event
environment = Environment.get_or_create(
project=project,
name=environment,
)
group_environment, is_new_group_environment = GroupEnvironment.get_or_create(
group_id=group.id,
environment_id=environment.id,
defaults={
'first_release_id': release.id if release else None,
},
)
if release:
ReleaseEnvironment.get_or_create(
project=project,
release=release,
environment=environment,
datetime=date,
)
ReleaseProjectEnvironment.get_or_create(
project=project,
release=release,
environment=environment,
datetime=date,
)
grouprelease = GroupRelease.get_or_create(
group=group,
release=release,
environment=environment,
datetime=date,
)
counters = [
(tsdb.models.group, group.id),
(tsdb.models.project, project.id),
]
if release:
counters.append((tsdb.models.release, release.id))
tsdb.incr_multi(counters, timestamp=event.datetime, environment_id=environment.id)
frequencies = [
# (tsdb.models.frequent_projects_by_organization, {
# project.organization_id: {
# project.id: 1,
# },
# }),
# (tsdb.models.frequent_issues_by_project, {
# project.id: {
# group.id: 1,
# },
# })
(tsdb.models.frequent_environments_by_group, {
group.id: {
environment.id: 1,
},
})
]
if release:
frequencies.append(
(tsdb.models.frequent_releases_by_group, {
group.id: {
grouprelease.id: 1,
},
})
)
tsdb.record_frequency_multi(frequencies, timestamp=event.datetime)
UserReport.objects.filter(
project=project,
event_id=event_id,
).update(
group=group,
environment=environment,
)
# save the event unless its been sampled
if not is_sample:
try:
with transaction.atomic(using=router.db_for_write(Event)):
event.save()
except IntegrityError:
self.logger.info(
'duplicate.found',
exc_info=True,
extra={
'event_uuid': event_id,
'project_id': project.id,
'group_id': group.id,
'model': Event.__name__,
}
)
return event
index_event_tags.delay(
organization_id=project.organization_id,
project_id=project.id,
group_id=group.id,
environment_id=environment.id,
event_id=event.id,
tags=tags,
date_added=event.datetime,
)
if event_user:
tsdb.record_multi(
(
(tsdb.models.users_affected_by_group, group.id, (event_user.tag_value, )),
(tsdb.models.users_affected_by_project, project.id, (event_user.tag_value, )),
),
timestamp=event.datetime,
environment_id=environment.id,
)
if release:
if is_new:
buffer.incr(
ReleaseProject, {'new_groups': 1}, {
'release_id': release.id,
'project_id': project.id,
}
)
if is_new_group_environment:
buffer.incr(
ReleaseProjectEnvironment, {'new_issues_count': 1}, {
'project_id': project.id,
'release_id': release.id,
'environment_id': environment.id,
}
)
safe_execute(Group.objects.add_tags, group, environment, tags, _with_transaction=False)
if not raw:
if not project.first_event:
project.update(first_event=date)
first_event_received.send_robust(project=project, group=group, sender=Project)
eventstream.insert(
group=group,
event=event,
is_new=is_new,
is_sample=is_sample,
is_regression=is_regression,
is_new_group_environment=is_new_group_environment,
primary_hash=hashes[0],
# We are choosing to skip consuming the event back
# in the eventstream if it's flagged as raw.
# This means that we want to publish the event
# through the event stream, but we don't care
# about post processing and handling the commit.
skip_consume=raw,
)
metrics.timing(
'events.latency',
received_timestamp - recorded_timestamp,
tags={
'project_id': project.id,
},
)
return event
def save(self, project, raw=False):
from sentry.tasks.post_process import index_event_tags
data = self.data
project = Project.objects.get_from_cache(id=project)
# Check to make sure we're not about to do a bunch of work that's
# already been done if we've processed an event with this ID. (This
# isn't a perfect solution -- this doesn't handle ``EventMapping`` and
# there's a race condition between here and when the event is actually
# saved, but it's an improvement. See GH-7677.)
try:
event = Event.objects.get(
project_id=project.id,
event_id=data['event_id'],
)
except Event.DoesNotExist:
pass
else:
self.logger.info('duplicate.found',
exc_info=True,
extra={
'event_uuid': data['event_id'],
'project_id': project.id,
'model': Event.__name__,
})
return event
# First we pull out our top-level (non-data attr) kwargs
event_id = data.pop('event_id')
level = data.pop('level')
culprit = data.pop('transaction', None)
if not culprit:
culprit = data.pop('culprit', None)
logger_name = data.pop('logger', None)
server_name = data.pop('server_name', None)
site = data.pop('site', None)
checksum = data.pop('checksum', None)
fingerprint = data.pop('fingerprint', None)
platform = data.pop('platform', None)
release = data.pop('release', None)
dist = data.pop('dist', None)
environment = data.pop('environment', None)
# unused
time_spent = data.pop('time_spent', None)
message = data.pop('message', '')
if not culprit:
# if we generate an implicit culprit, lets not call it a
# transaction
transaction_name = None
culprit = generate_culprit(data, platform=platform)
else:
transaction_name = culprit
culprit = force_text(culprit)
recorded_timestamp = data.pop('timestamp')
date = datetime.fromtimestamp(recorded_timestamp)
date = date.replace(tzinfo=timezone.utc)
kwargs = {
'platform': platform,
}
event = Event(project_id=project.id,
event_id=event_id,
data=data,
time_spent=time_spent,
datetime=date,
**kwargs)
event._project_cache = project
# convert this to a dict to ensure we're only storing one value per key
# as most parts of Sentry dont currently play well with multiple values
tags = dict(data.get('tags') or [])
tags['level'] = LOG_LEVELS[level]
if logger_name:
tags['logger'] = logger_name
if server_name:
tags['server_name'] = server_name
if site:
tags['site'] = site
if environment:
tags['environment'] = environment
if transaction_name:
tags['transaction'] = transaction_name
if release:
# dont allow a conflicting 'release' tag
if 'release' in tags:
del tags['release']
release = Release.get_or_create(
project=project,
version=release,
date_added=date,
)
tags['sentry:release'] = release.version
if dist and release:
dist = release.add_dist(dist, date)
tags['sentry:dist'] = dist.name
else:
dist = None
event_user = self._get_event_user(project, data)
if event_user:
# dont allow a conflicting 'user' tag
if 'user' in tags:
del tags['user']
tags['sentry:user'] = event_user.tag_value
# At this point we want to normalize the in_app values in case the
# clients did not set this appropriately so far.
normalize_in_app(data)
for plugin in plugins.for_project(project, version=None):
added_tags = safe_execute(plugin.get_tags,
event,
_with_transaction=False)
if added_tags:
# plugins should not override user provided tags
for key, value in added_tags:
tags.setdefault(key, value)
for path, iface in six.iteritems(event.interfaces):
for k, v in iface.iter_tags():
tags[k] = v
# Get rid of ephemeral interface data
if iface.ephemeral:
data.pop(iface.get_path(), None)
# tags are stored as a tuple
tags = tags.items()
data['tags'] = tags
data['fingerprint'] = fingerprint or ['{{ default }}']
# prioritize fingerprint over checksum as its likely the client defaulted
# a checksum whereas the fingerprint was explicit
if fingerprint:
hashes = [
md5_from_hash(h)
for h in get_hashes_from_fingerprint(event, fingerprint)
]
elif checksum:
if HASH_RE.match(checksum):
hashes = [checksum]
else:
hashes = [md5_from_hash([checksum]), checksum]
data['checksum'] = checksum
else:
hashes = [md5_from_hash(h) for h in get_hashes_for_event(event)]
# TODO(dcramer): temp workaround for complexity
data['message'] = message
event_type = eventtypes.get(data.get('type', 'default'))(data)
event_metadata = event_type.get_metadata()
# TODO(dcramer): temp workaround for complexity
del data['message']
data['type'] = event_type.key
data['metadata'] = event_metadata
# index components into ``Event.message``
# See GH-3248
if event_type.key != 'default':
if 'sentry.interfaces.Message' in data and \
data['sentry.interfaces.Message']['message'] != message:
message = u'{} {}'.format(
message,
data['sentry.interfaces.Message']['message'],
)
if not message:
message = ''
elif not isinstance(message, six.string_types):
message = force_text(message)
for value in six.itervalues(event_metadata):
value_u = force_text(value, errors='replace')
if value_u not in message:
message = u'{} {}'.format(message, value_u)
if culprit and culprit not in message:
culprit_u = force_text(culprit, errors='replace')
message = u'{} {}'.format(message, culprit_u)
message = trim(message.strip(), settings.SENTRY_MAX_MESSAGE_LENGTH)
event.message = message
kwargs['message'] = message
received_timestamp = event.data.get('received') or float(
event.datetime.strftime('%s'))
group_kwargs = kwargs.copy()
group_kwargs.update({
'culprit': culprit,
'logger': logger_name,
'level': level,
'last_seen': date,
'first_seen': date,
'active_at': date,
'data': {
'last_received': received_timestamp,
'type': event_type.key,
# we cache the events metadata on the group to ensure its
# accessible in the stream
'metadata': event_metadata,
},
})
if release:
group_kwargs['first_release'] = release
try:
group, is_new, is_regression, is_sample = self._save_aggregate(
event=event, hashes=hashes, release=release, **group_kwargs)
except HashDiscarded:
event_discarded.send_robust(
project=project,
sender=EventManager,
)
metrics.incr(
'events.discarded',
skip_internal=True,
tags={
'organization_id': project.organization_id,
'platform': platform,
},
)
raise
else:
event_saved.send_robust(
project=project,
sender=EventManager,
)
event.group = group
# store a reference to the group id to guarantee validation of isolation
event.data.bind_ref(event)
# When an event was sampled, the canonical source of truth
# is the EventMapping table since we aren't going to be writing out an actual
# Event row. Otherwise, if the Event isn't being sampled, we can safely
# rely on the Event table itself as the source of truth and ignore
# EventMapping since it's redundant information.
if is_sample:
try:
with transaction.atomic(
using=router.db_for_write(EventMapping)):
EventMapping.objects.create(project=project,
group=group,
event_id=event_id)
except IntegrityError:
self.logger.info('duplicate.found',
exc_info=True,
extra={
'event_uuid': event_id,
'project_id': project.id,
'group_id': group.id,
'model': EventMapping.__name__,
})
return event
environment = Environment.get_or_create(
project=project,
name=environment,
)
group_environment, is_new_group_environment = GroupEnvironment.get_or_create(
group_id=group.id,
environment_id=environment.id,
defaults={
'first_release_id': release.id if release else None,
},
)
if release:
ReleaseEnvironment.get_or_create(
project=project,
release=release,
environment=environment,
datetime=date,
)
ReleaseProjectEnvironment.get_or_create(
project=project,
release=release,
environment=environment,
datetime=date,
)
grouprelease = GroupRelease.get_or_create(
group=group,
release=release,
environment=environment,
datetime=date,
)
counters = [
(tsdb.models.group, group.id),
(tsdb.models.project, project.id),
]
if release:
counters.append((tsdb.models.release, release.id))
tsdb.incr_multi(counters,
timestamp=event.datetime,
environment_id=environment.id)
frequencies = [
# (tsdb.models.frequent_projects_by_organization, {
# project.organization_id: {
# project.id: 1,
# },
# }),
# (tsdb.models.frequent_issues_by_project, {
# project.id: {
# group.id: 1,
# },
# })
(tsdb.models.frequent_environments_by_group, {
group.id: {
environment.id: 1,
},
})
]
if release:
frequencies.append((tsdb.models.frequent_releases_by_group, {
group.id: {
grouprelease.id: 1,
},
}))
tsdb.record_frequency_multi(frequencies, timestamp=event.datetime)
UserReport.objects.filter(
project=project,
event_id=event_id,
).update(
group=group,
environment=environment,
)
# save the event unless its been sampled
if not is_sample:
try:
with transaction.atomic(using=router.db_for_write(Event)):
event.save()
except IntegrityError:
self.logger.info('duplicate.found',
exc_info=True,
extra={
'event_uuid': event_id,
'project_id': project.id,
'group_id': group.id,
'model': Event.__name__,
})
return event
index_event_tags.delay(
organization_id=project.organization_id,
project_id=project.id,
group_id=group.id,
environment_id=environment.id,
event_id=event.id,
tags=tags,
date_added=event.datetime,
)
if event_user:
tsdb.record_multi(
(
(tsdb.models.users_affected_by_group, group.id,
(event_user.tag_value, )),
(tsdb.models.users_affected_by_project, project.id,
(event_user.tag_value, )),
),
timestamp=event.datetime,
environment_id=environment.id,
)
if release:
if is_new:
buffer.incr(ReleaseProject, {'new_groups': 1}, {
'release_id': release.id,
'project_id': project.id,
})
if is_new_group_environment:
buffer.incr(ReleaseProjectEnvironment, {'new_issues_count': 1},
{
'project_id': project.id,
'release_id': release.id,
'environment_id': environment.id,
})
safe_execute(Group.objects.add_tags,
group,
environment,
tags,
_with_transaction=False)
if not raw:
if not project.first_event:
project.update(first_event=date)
first_event_received.send(project=project,
group=group,
sender=Project)
post_process_group.delay(
group=group,
event=event,
is_new=is_new,
is_sample=is_sample,
is_regression=is_regression,
is_new_group_environment=is_new_group_environment,
primary_hash=hashes[0],
)
else:
self.logger.info('post_process.skip.raw_event',
extra={'event_id': event.id})
metrics.timing(
'events.latency',
received_timestamp - recorded_timestamp,
tags={
'project_id': project.id,
},
)
return event
def save(self, project, raw=False):
project = Project.objects.get_from_cache(id=project)
data = self.data.copy()
# First we pull out our top-level (non-data attr) kwargs
event_id = data.pop('event_id')
message = data.pop('message')
level = data.pop('level')
culprit = data.pop('culprit', None)
time_spent = data.pop('time_spent', None)
logger_name = data.pop('logger', None)
server_name = data.pop('server_name', None)
site = data.pop('site', None)
checksum = data.pop('checksum', None)
fingerprint = data.pop('fingerprint', None)
platform = data.pop('platform', None)
release = data.pop('release', None)
environment = data.pop('environment', None)
if not culprit:
culprit = generate_culprit(data)
date = datetime.fromtimestamp(data.pop('timestamp'))
date = date.replace(tzinfo=timezone.utc)
kwargs = {
'message': message,
'platform': platform,
}
event = Event(project_id=project.id,
event_id=event_id,
data=data,
time_spent=time_spent,
datetime=date,
**kwargs)
tags = data.get('tags') or []
tags.append(('level', LOG_LEVELS[level]))
if logger_name:
tags.append(('logger', logger_name))
if server_name:
tags.append(('server_name', server_name))
if site:
tags.append(('site', site))
if release:
# TODO(dcramer): we should ensure we create Release objects
tags.append(('sentry:release', release))
if environment:
tags.append(('environment', environment))
for plugin in plugins.for_project(project, version=None):
added_tags = safe_execute(plugin.get_tags,
event,
_with_transaction=False)
if added_tags:
tags.extend(added_tags)
event_user = self._get_event_user(project, data)
if event_user:
tags.append(('sentry:user', event_user.tag_value))
# XXX(dcramer): we're relying on mutation of the data object to ensure
# this propagates into Event
data['tags'] = tags
data['fingerprint'] = fingerprint or ['{{ default }}']
# prioritize fingerprint over checksum as its likely the client defaulted
# a checksum whereas the fingerprint was explicit
if fingerprint:
hashes = map(md5_from_hash,
get_hashes_from_fingerprint(event, fingerprint))
elif checksum:
hashes = [checksum]
else:
hashes = map(md5_from_hash, get_hashes_for_event(event))
group_kwargs = kwargs.copy()
group_kwargs.update({
'culprit': culprit,
'logger': logger_name,
'level': level,
'last_seen': date,
'first_seen': date,
'time_spent_total': time_spent or 0,
'time_spent_count': time_spent and 1 or 0,
})
if release:
release = Release.get_or_create(
project=project,
version=release,
date_added=date,
)
group_kwargs['first_release'] = release
group, is_new, is_regression, is_sample = self._save_aggregate(
event=event, hashes=hashes, release=release, **group_kwargs)
event.group = group
# store a reference to the group id to guarantee validation of isolation
event.data.bind_ref(event)
try:
with transaction.atomic(using=router.db_for_write(EventMapping)):
EventMapping.objects.create(project=project,
group=group,
event_id=event_id)
except IntegrityError:
self.logger.info('Duplicate EventMapping found for event_id=%s',
event_id,
exc_info=True)
return event
UserReport.objects.filter(
project=project,
event_id=event_id,
).update(group=group)
# save the event unless its been sampled
if not is_sample:
try:
with transaction.atomic(using=router.db_for_write(Event)):
event.save()
except IntegrityError:
self.logger.info('Duplicate Event found for event_id=%s',
event_id,
exc_info=True)
return event
if event_user:
tsdb.record_multi((
(tsdb.models.users_affected_by_group, group.id,
(event_user.tag_value, )),
(tsdb.models.users_affected_by_project, project.id,
(event_user.tag_value, )),
),
timestamp=event.datetime)
if is_new and release:
buffer.incr(Release, {'new_groups': 1}, {
'id': release.id,
})
safe_execute(Group.objects.add_tags,
group,
tags,
_with_transaction=False)
if not raw:
if not project.first_event:
project.update(first_event=date)
post_process_group.delay(
group=group,
event=event,
is_new=is_new,
is_sample=is_sample,
is_regression=is_regression,
)
else:
self.logger.info(
'Raw event passed; skipping post process for event_id=%s',
event_id)
# TODO: move this to the queue
if is_regression and not raw:
regression_signal.send_robust(sender=Group, instance=group)
return event
def save(self, project, raw=False):
from sentry.tasks.post_process import index_event_tags
project = Project.objects.get_from_cache(id=project)
data = self.data.copy()
# First we pull out our top-level (non-data attr) kwargs
event_id = data.pop("event_id")
message = data.pop("message")
level = data.pop("level")
culprit = data.pop("culprit", None)
logger_name = data.pop("logger", None)
server_name = data.pop("server_name", None)
site = data.pop("site", None)
checksum = data.pop("checksum", None)
fingerprint = data.pop("fingerprint", None)
platform = data.pop("platform", None)
release = data.pop("release", None)
environment = data.pop("environment", None)
# unused
time_spent = data.pop("time_spent", None)
if not culprit:
culprit = generate_culprit(data, platform=platform)
date = datetime.fromtimestamp(data.pop("timestamp"))
date = date.replace(tzinfo=timezone.utc)
kwargs = {"message": message, "platform": platform}
event = Event(
project_id=project.id, event_id=event_id, data=data, time_spent=time_spent, datetime=date, **kwargs
)
tags = data.get("tags") or []
tags.append(("level", LOG_LEVELS[level]))
if logger_name:
tags.append(("logger", logger_name))
if server_name:
tags.append(("server_name", server_name))
if site:
tags.append(("site", site))
if release:
# TODO(dcramer): we should ensure we create Release objects
tags.append(("sentry:release", release))
if environment:
tags.append(("environment", environment))
for plugin in plugins.for_project(project, version=None):
added_tags = safe_execute(plugin.get_tags, event, _with_transaction=False)
if added_tags:
tags.extend(added_tags)
event_user = self._get_event_user(project, data)
if event_user:
tags.append(("sentry:user", event_user.tag_value))
# XXX(dcramer): we're relying on mutation of the data object to ensure
# this propagates into Event
data["tags"] = tags
data["fingerprint"] = fingerprint or ["{{ default }}"]
# Get rid of ephemeral interface data
for interface_class, _ in iter_interfaces():
interface = interface_class()
if interface.ephemeral:
data.pop(interface.get_path(), None)
# prioritize fingerprint over checksum as its likely the client defaulted
# a checksum whereas the fingerprint was explicit
if fingerprint:
hashes = map(md5_from_hash, get_hashes_from_fingerprint(event, fingerprint))
elif checksum:
hashes = [checksum]
else:
hashes = map(md5_from_hash, get_hashes_for_event(event))
# TODO(dcramer): temp workaround for complexity
data["message"] = message
event_type = eventtypes.get(data.get("type", "default"))(data)
group_kwargs = kwargs.copy()
group_kwargs.update(
{
"culprit": culprit,
"logger": logger_name,
"level": level,
"last_seen": date,
"first_seen": date,
"data": {
"last_received": event.data.get("received") or float(event.datetime.strftime("%s")),
"type": event_type.key,
# we cache the events metadata on the group to ensure its
# accessible in the stream
"metadata": event_type.get_metadata(),
},
}
)
# TODO(dcramer): temp workaround for complexity
del data["message"]
if release:
release = Release.get_or_create(project=project, version=release, date_added=date)
group_kwargs["first_release"] = release
group, is_new, is_regression, is_sample = self._save_aggregate(
event=event, hashes=hashes, release=release, **group_kwargs
)
event.group = group
# store a reference to the group id to guarantee validation of isolation
event.data.bind_ref(event)
try:
with transaction.atomic(using=router.db_for_write(EventMapping)):
EventMapping.objects.create(project=project, group=group, event_id=event_id)
except IntegrityError:
self.logger.info("Duplicate EventMapping found for event_id=%s", event_id, exc_info=True)
return event
UserReport.objects.filter(project=project, event_id=event_id).update(group=group)
# save the event unless its been sampled
if not is_sample:
try:
with transaction.atomic(using=router.db_for_write(Event)):
event.save()
except IntegrityError:
self.logger.info("Duplicate Event found for event_id=%s", event_id, exc_info=True)
return event
index_event_tags.delay(project_id=project.id, event_id=event.id, tags=tags)
if event_user:
tsdb.record_multi(
(
(tsdb.models.users_affected_by_group, group.id, (event_user.tag_value,)),
(tsdb.models.users_affected_by_project, project.id, (event_user.tag_value,)),
),
timestamp=event.datetime,
)
if is_new and release:
buffer.incr(Release, {"new_groups": 1}, {"id": release.id})
safe_execute(Group.objects.add_tags, group, tags, _with_transaction=False)
if not raw:
if not project.first_event:
project.update(first_event=date)
first_event_received.send(project=project, group=group, sender=Project)
post_process_group.delay(
group=group, event=event, is_new=is_new, is_sample=is_sample, is_regression=is_regression
)
else:
self.logger.info("Raw event passed; skipping post process for event_id=%s", event_id)
# TODO: move this to the queue
if is_regression and not raw:
regression_signal.send_robust(sender=Group, instance=group)
return event
def get_event_file_committers(project, event, frame_limit=25):
# populate event data
Event.objects.bind_nodes([event], 'data')
group = Group.objects.get(id=event.group_id)
first_release_version = group.get_first_release()
if not first_release_version:
raise Release.DoesNotExist
releases = Release.get_closest_releases(project, first_release_version)
if not releases:
raise Release.DoesNotExist
commits = _get_commits(releases)
if not commits:
raise Commit.DoesNotExist
frames = _get_frame_paths(event)
app_frames = [frame for frame in frames if frame['in_app']][-frame_limit:]
if not app_frames:
app_frames = [frame for frame in frames][-frame_limit:]
# TODO(maxbittker) return this set instead of annotated frames
# XXX(dcramer): frames may not define a filepath. For example, in Java its common
# to only have a module name/path
path_set = {
f
for f in (frame.get('filename') or frame.get('abs_path')
for frame in app_frames) if f
}
file_changes = []
if path_set:
file_changes = _get_commit_file_changes(commits, path_set)
commit_path_matches = {
path: _match_commits_path(file_changes, path)
for path in path_set
}
annotated_frames = [{
'frame':
frame,
'commits':
commit_path_matches.get(
frame.get('filename') or frame.get('abs_path')) or []
} for frame in app_frames]
relevant_commits = list({
match
for match in commit_path_matches
for match in commit_path_matches[match]
})
committers = _get_committers(annotated_frames, relevant_commits)
metrics.incr('feature.owners.has-committers',
instance='hit' if committers else 'miss')
return committers
def setUp(self):
self.project = self.create_project()
self.release = Release.get_or_create(self.project, '1.0')
self.environment1 = Environment.get_or_create(self.project, 'prod')
self.environment2 = Environment.get_or_create(self.project, 'staging')
self.timestamp = 1403007314
def merge_to(from_org, to_org):
from sentry.models import (
ApiKey, AuditLogEntry, Commit, OrganizationMember,
OrganizationMemberTeam, Project, Release, ReleaseCommit,
ReleaseEnvironment, ReleaseFile, Repository, Team
)
for from_member in OrganizationMember.objects.filter(organization=from_org):
try:
to_member = OrganizationMember.objects.get(
organization=to_org,
user=from_member.user,
)
except OrganizationMember.DoesNotExist:
from_member.update(organization=to_org)
to_member = from_member
else:
qs = OrganizationMemberTeam.objects.filter(
organizationmember=from_member,
is_active=True,
).select_related()
for omt in qs:
OrganizationMemberTeam.objects.create_or_update(
organizationmember=to_member,
team=omt.team,
defaults={
'is_active': True,
},
)
for team in Team.objects.filter(organization=from_org):
try:
with transaction.atomic():
team.update(organization=to_org)
except IntegrityError:
slugify_instance(team, team.name, organization=to_org)
team.update(
organization=to_org,
slug=team.slug,
)
for project in Project.objects.filter(organization=from_org):
try:
with transaction.atomic():
project.update(organization=to_org)
except IntegrityError:
slugify_instance(project, project.name, organization=to_org)
project.update(
organization=to_org,
slug=project.slug,
)
# TODO(jess): update this when adding unique constraint
# on version, organization for releases
for release in Release.objects.filter(organization=from_org):
try:
to_release = Release.objects.get(
version=release.version,
organization=to_org
)
except Release.DoesNotExist:
Release.objects.filter(
id=release.id
).update(organization=to_org)
else:
Release.merge(to_release, [release])
for model in (ApiKey, AuditLogEntry, ReleaseFile):
model.objects.filter(
organization=from_org,
).update(organization=to_org)
for model in (Commit, ReleaseCommit, ReleaseEnvironment, Repository):
model.objects.filter(
organization_id=from_org.id,
).update(organization_id=to_org.id)
def save(self, project, raw=False):
from sentry.tasks.post_process import index_event_tags
project = Project.objects.get_from_cache(id=project)
data = self.data.copy()
# First we pull out our top-level (non-data attr) kwargs
event_id = data.pop('event_id')
level = data.pop('level')
culprit = data.pop('transaction', None)
if not culprit:
culprit = data.pop('culprit', None)
logger_name = data.pop('logger', None)
server_name = data.pop('server_name', None)
site = data.pop('site', None)
checksum = data.pop('checksum', None)
fingerprint = data.pop('fingerprint', None)
platform = data.pop('platform', None)
release = data.pop('release', None)
dist = data.pop('dist', None)
environment = data.pop('environment', None)
# unused
time_spent = data.pop('time_spent', None)
message = data.pop('message', '')
if not culprit:
# if we generate an implicit culprit, lets not call it a
# transaction
transaction_name = None
culprit = generate_culprit(data, platform=platform)
else:
transaction_name = culprit
date = datetime.fromtimestamp(data.pop('timestamp'))
date = date.replace(tzinfo=timezone.utc)
kwargs = {
'platform': platform,
}
event = Event(
project_id=project.id,
event_id=event_id,
data=data,
time_spent=time_spent,
datetime=date,
**kwargs
)
event._project_cache = project
# convert this to a dict to ensure we're only storing one value per key
# as most parts of Sentry dont currently play well with multiple values
tags = dict(data.get('tags') or [])
tags['level'] = LOG_LEVELS[level]
if logger_name:
tags['logger'] = logger_name
if server_name:
tags['server_name'] = server_name
if site:
tags['site'] = site
if environment:
tags['environment'] = environment
if transaction_name:
tags['transaction'] = transaction_name
if release:
# dont allow a conflicting 'release' tag
if 'release' in tags:
del tags['release']
release = Release.get_or_create(
project=project,
version=release,
date_added=date,
)
tags['sentry:release'] = release.version
if dist and release:
dist = release.add_dist(dist, date)
tags['sentry:dist'] = dist.name
else:
dist = None
event_user = self._get_event_user(project, data)
if event_user:
# dont allow a conflicting 'user' tag
if 'user' in tags:
del tags['user']
tags['sentry:user'] = event_user.tag_value
# At this point we want to normalize the in_app values in case the
# clients did not set this appropriately so far.
normalize_in_app(data)
for plugin in plugins.for_project(project, version=None):
added_tags = safe_execute(plugin.get_tags, event, _with_transaction=False)
if added_tags:
# plugins should not override user provided tags
for key, value in added_tags:
tags.setdefault(key, value)
# tags are stored as a tuple
tags = tags.items()
# XXX(dcramer): we're relying on mutation of the data object to ensure
# this propagates into Event
data['tags'] = tags
data['fingerprint'] = fingerprint or ['{{ default }}']
for path, iface in six.iteritems(event.interfaces):
data['tags'].extend(iface.iter_tags())
# Get rid of ephemeral interface data
if iface.ephemeral:
data.pop(iface.get_path(), None)
# prioritize fingerprint over checksum as its likely the client defaulted
# a checksum whereas the fingerprint was explicit
if fingerprint:
hashes = [md5_from_hash(h) for h in get_hashes_from_fingerprint(event, fingerprint)]
elif checksum:
hashes = [checksum]
data['checksum'] = checksum
else:
hashes = [md5_from_hash(h) for h in get_hashes_for_event(event)]
# TODO(dcramer): temp workaround for complexity
data['message'] = message
event_type = eventtypes.get(data.get('type', 'default'))(data)
event_metadata = event_type.get_metadata()
# TODO(dcramer): temp workaround for complexity
del data['message']
data['type'] = event_type.key
data['metadata'] = event_metadata
# index components into ``Event.message``
# See GH-3248
if event_type.key != 'default':
if 'sentry.interfaces.Message' in data and \
data['sentry.interfaces.Message']['message'] != message:
message = u'{} {}'.format(
message,
data['sentry.interfaces.Message']['message'],
)
if not message:
message = ''
elif not isinstance(message, six.string_types):
message = force_text(message)
for value in six.itervalues(event_metadata):
value_u = force_text(value, errors='replace')
if value_u not in message:
message = u'{} {}'.format(message, value_u)
if culprit and culprit not in message:
culprit_u = force_text(culprit, errors='replace')
message = u'{} {}'.format(message, culprit_u)
message = trim(message.strip(), settings.SENTRY_MAX_MESSAGE_LENGTH)
event.message = message
kwargs['message'] = message
group_kwargs = kwargs.copy()
group_kwargs.update(
{
'culprit': culprit,
'logger': logger_name,
'level': level,
'last_seen': date,
'first_seen': date,
'active_at': date,
'data': {
'last_received':
event.data.get('received') or float(event.datetime.strftime('%s')),
'type':
event_type.key,
# we cache the events metadata on the group to ensure its
# accessible in the stream
'metadata':
event_metadata,
},
}
)
if release:
group_kwargs['first_release'] = release
group, is_new, is_regression, is_sample = self._save_aggregate(
event=event, hashes=hashes, release=release, **group_kwargs
)
event.group = group
# store a reference to the group id to guarantee validation of isolation
event.data.bind_ref(event)
try:
with transaction.atomic(using=router.db_for_write(EventMapping)):
EventMapping.objects.create(project=project, group=group, event_id=event_id)
except IntegrityError:
self.logger.info(
'duplicate.found',
exc_info=True,
extra={
'event_uuid': event_id,
'project_id': project.id,
'group_id': group.id,
'model': EventMapping.__name__,
}
)
return event
environment = Environment.get_or_create(
project=project,
name=environment,
)
if release:
ReleaseEnvironment.get_or_create(
project=project,
release=release,
environment=environment,
datetime=date,
)
grouprelease = GroupRelease.get_or_create(
group=group,
release=release,
environment=environment,
datetime=date,
)
counters = [
(tsdb.models.group, group.id),
(tsdb.models.project, project.id),
]
if release:
counters.append((tsdb.models.release, release.id))
tsdb.incr_multi(counters, timestamp=event.datetime)
frequencies = [
# (tsdb.models.frequent_projects_by_organization, {
# project.organization_id: {
# project.id: 1,
# },
# }),
# (tsdb.models.frequent_issues_by_project, {
# project.id: {
# group.id: 1,
# },
# })
(tsdb.models.frequent_environments_by_group, {
group.id: {
environment.id: 1,
},
})
]
if release:
frequencies.append(
(tsdb.models.frequent_releases_by_group, {
group.id: {
grouprelease.id: 1,
},
})
)
tsdb.record_frequency_multi(frequencies, timestamp=event.datetime)
UserReport.objects.filter(
project=project,
event_id=event_id,
).update(group=group)
# save the event unless its been sampled
if not is_sample:
try:
with transaction.atomic(using=router.db_for_write(Event)):
event.save()
except IntegrityError:
self.logger.info(
'duplicate.found',
exc_info=True,
extra={
'event_uuid': event_id,
'project_id': project.id,
'group_id': group.id,
'model': Event.__name__,
}
)
return event
index_event_tags.delay(
organization_id=project.organization_id,
project_id=project.id,
group_id=group.id,
event_id=event.id,
tags=tags,
)
if event_user:
tsdb.record_multi(
(
(tsdb.models.users_affected_by_group, group.id, (event_user.tag_value, )),
(tsdb.models.users_affected_by_project, project.id, (event_user.tag_value, )),
),
timestamp=event.datetime
)
if is_new and release:
buffer.incr(
ReleaseProject, {'new_groups': 1}, {
'release_id': release.id,
'project_id': project.id,
}
)
safe_execute(Group.objects.add_tags, group, tags, _with_transaction=False)
if not raw:
if not project.first_event:
project.update(first_event=date)
first_event_received.send(project=project, group=group, sender=Project)
post_process_group.delay(
group=group,
event=event,
is_new=is_new,
is_sample=is_sample,
is_regression=is_regression,
)
else:
self.logger.info('post_process.skip.raw_event', extra={'event_id': event.id})
# TODO: move this to the queue
if is_regression and not raw:
regression_signal.send_robust(sender=Group, instance=group)
return event
def create_message_event(template, parameters, environment, release):
i = next(sequence)
event_id = uuid.UUID(
fields=(i, 0x0, 0x1000, 0x80, 0x80, 0x808080808080, ),
).hex
tags = [['color', next(tag_values)]]
if environment:
tags.append(['environment', environment])
if release:
tags.append(['sentry:release', release])
event = Event.objects.create(
project_id=project.id,
group_id=source.id,
event_id=event_id,
message='%s' % (id, ),
datetime=now + shift(i),
data={
'environment': environment,
'type': 'default',
'metadata': {
'title': template % parameters,
},
'logentry': {
'message': template,
'params': parameters,
'formatted': template % parameters,
},
'user': next(user_values),
'tags': tags,
},
)
with self.tasks():
Group.objects.add_tags(
source,
Environment.objects.get(
organization_id=project.organization_id,
name=environment
),
tags=event.get_tags(),
)
EventMapping.objects.create(
project_id=project.id,
group_id=source.id,
event_id=event_id,
date_added=event.datetime,
)
UserReport.objects.create(
project_id=project.id,
group_id=source.id,
event_id=event_id,
name='Log Hat',
email='*****@*****.**',
comments='Quack',
)
if release:
Release.get_or_create(
project=project,
version=event.get_tag('sentry:release'),
date_added=event.datetime,
)
features.record([event])
return event
def post(self, request, project):
"""
Create a New Release
````````````````````
Create a new release and/or associate a project with a release.
Release versions that are the same across multiple projects
within an Organization will be treated as the same release in Sentry.
Releases are used by Sentry to improve its error reporting abilities
by correlating first seen events with the release that might have
introduced the problem.
Releases are also necessary for sourcemaps and other debug features
that require manual upload for functioning well.
:pparam string organization_slug: the slug of the organization the
release belongs to.
:pparam string project_slug: the slug of the project to create a
release for.
:param string version: a version identifier for this release. Can
be a version number, a commit hash etc.
:param string ref: an optional commit reference. This is useful if
a tagged version has been provided.
:param url url: a URL that points to the release. This can be the
path to an online interface to the sourcecode
for instance.
:param datetime dateStarted: an optional date that indicates when the
release process started.
:param datetime dateReleased: an optional date that indicates when
the release went live. If not provided
the current time is assumed.
:auth: required
"""
serializer = ReleaseSerializer(data=request.DATA)
if serializer.is_valid():
result = serializer.object
# release creation is idempotent to simplify user
# experiences
release = Release.objects.filter(
organization_id=project.organization_id,
version=result['version'],
projects=project).first()
created = False
if release:
was_released = bool(release.date_released)
else:
release = Release.objects.filter(
organization_id=project.organization_id,
version=result['version'],
).first()
if not release:
lock_key = Release.get_lock_key(project.organization_id,
result['version'])
lock = locks.get(lock_key, duration=5)
with TimedRetryPolicy(10)(lock.acquire):
try:
release, created = Release.objects.get(
version=result['version'],
organization_id=project.organization_id), False
except Release.DoesNotExist:
release, created = Release.objects.create(
organization_id=project.organization_id,
version=result['version'],
ref=result.get('ref'),
url=result.get('url'),
owner=result.get('owner'),
date_started=result.get('dateStarted'),
date_released=result.get('dateReleased'),
), True
was_released = False
try:
with transaction.atomic():
ReleaseProject.objects.create(project=project,
release=release)
created = True
except IntegrityError:
pass
commit_list = result.get('commits')
if commit_list:
hook = ReleaseHook(project)
# TODO(dcramer): handle errors with release payloads
hook.set_commits(release.version, commit_list)
if (not was_released and release.date_released):
activity = Activity.objects.create(
type=Activity.RELEASE,
project=project,
ident=result['version'],
data={'version': result['version']},
datetime=release.date_released,
)
activity.send_notification()
if not created:
# This is the closest status code that makes sense, and we want
# a unique 2xx response code so people can understand when
# behavior differs.
# 208 Already Reported (WebDAV; RFC 5842)
status = 208
else:
status = 201
return Response(serialize(release, request.user), status=status)
return Response(serializer.errors, status=400)
def merge_to(from_org, to_org):
from sentry.models import (
ApiKey,
AuditLogEntry,
AuthProvider,
Commit,
OrganizationAvatar,
OrganizationIntegration,
OrganizationMember,
OrganizationMemberTeam,
Project,
Release,
ReleaseCommit,
ReleaseEnvironment,
ReleaseFile,
ReleaseHeadCommit,
Repository,
Team,
Environment,
)
for from_member in OrganizationMember.objects.filter(
organization=from_org, user__isnull=False
):
logger = logging.getLogger('sentry.merge')
try:
to_member = OrganizationMember.objects.get(
organization=to_org,
user=from_member.user,
)
except OrganizationMember.DoesNotExist:
from_member.update(organization=to_org)
to_member = from_member
else:
qs = OrganizationMemberTeam.objects.filter(
organizationmember=from_member,
is_active=True,
).select_related()
for omt in qs:
OrganizationMemberTeam.objects.create_or_update(
organizationmember=to_member,
team=omt.team,
defaults={
'is_active': True,
},
)
logger.info('user.migrate', extra={
'instance_id': from_member.id,
'new_member_id': to_member.id,
'from_organization_id': from_org.id,
'to_organization_id': to_org.id,
})
for from_team in Team.objects.filter(organization=from_org):
try:
with transaction.atomic():
from_team.update(organization=to_org)
except IntegrityError:
slugify_instance(from_team, from_team.name, organization=to_org)
from_team.update(
organization=to_org,
slug=from_team.slug,
)
logger.info('team.migrate', extra={
'instance_id': from_team.id,
'new_slug': from_team.slug,
'from_organization_id': from_org.id,
'to_organization_id': to_org.id,
})
for from_project in Project.objects.filter(organization=from_org):
try:
with transaction.atomic():
from_project.update(organization=to_org)
except IntegrityError:
slugify_instance(
from_project,
from_project.name,
organization=to_org,
reserved=RESERVED_PROJECT_SLUGS)
from_project.update(
organization=to_org,
slug=from_project.slug,
)
logger.info('project.migrate', extra={
'instance_id': from_project.id,
'new_slug': from_project.slug,
'from_organization_id': from_org.id,
'to_organization_id': to_org.id,
})
# TODO(jess): update this when adding unique constraint
# on version, organization for releases
for from_release in Release.objects.filter(organization=from_org):
try:
to_release = Release.objects.get(version=from_release.version, organization=to_org)
except Release.DoesNotExist:
Release.objects.filter(id=from_release.id).update(organization=to_org)
else:
Release.merge(to_release, [from_release])
logger.info('release.migrate', extra={
'instance_id': from_release.id,
'from_organization_id': from_org.id,
'to_organization_id': to_org.id,
})
def do_update(queryset, params):
model_name = queryset.model.__name__.lower()
try:
with transaction.atomic():
queryset.update(**params)
except IntegrityError:
for instance in queryset:
try:
with transaction.atomic():
instance.update(**params)
except IntegrityError:
logger.info('{}.migrate-skipped'.format(model_name), extra={
'from_organization_id': from_org.id,
'to_organization_id': to_org.id,
})
else:
logger.info('{}.migrate'.format(model_name), extra={
'instance_id': instance.id,
'from_organization_id': from_org.id,
'to_organization_id': to_org.id,
})
else:
logger.info('{}.migrate'.format(model_name), extra={
'from_organization_id': from_org.id,
'to_organization_id': to_org.id,
})
INST_MODEL_LIST = (
AuthProvider, ApiKey, AuditLogEntry, OrganizationAvatar,
OrganizationIntegration, ReleaseEnvironment, ReleaseFile,
)
ATTR_MODEL_LIST = (
Commit, ReleaseCommit, ReleaseHeadCommit, Repository, Environment,
)
for model in INST_MODEL_LIST:
queryset = model.objects.filter(
organization=from_org,
)
do_update(queryset, {'organization': to_org})
for model in ATTR_MODEL_LIST:
queryset = model.objects.filter(
organization_id=from_org.id,
)
do_update(queryset, {'organization_id': to_org.id})
def test_simple(self):
org = self.create_organization()
commit = Commit.objects.create(organization_id=org.id, repository_id=5)
commit2 = Commit.objects.create(organization_id=org.id, repository_id=6)
# merge to
project = self.create_project(organization=org, name='foo')
environment = Environment.get_or_create(project=project, name='env1')
release = Release.objects.create(version='abcdabc', organization=org)
release.add_project(project)
release_commit = ReleaseCommit.objects.create(
organization_id=org.id, release=release, commit=commit, order=1
)
release_environment = ReleaseEnvironment.objects.create(
organization_id=org.id, project_id=project.id, release_id=release.id, environment_id=environment.id
)
release_project_environment = ReleaseProjectEnvironment.objects.create(
release_id=release.id, project_id=project.id, environment_id=environment.id
)
group_release = GroupRelease.objects.create(
project_id=project.id, release_id=release.id, group_id=1
)
group = self.create_group(project=project, first_release=release)
group_resolution = GroupResolution.objects.create(group=group, release=release)
# merge from #1
project2 = self.create_project(organization=org, name='bar')
environment2 = Environment.get_or_create(project=project2, name='env2')
release2 = Release.objects.create(version='bbbbbbb', organization=org)
release2.add_project(project2)
release_commit2 = ReleaseCommit.objects.create(
organization_id=org.id, release=release2, commit=commit, order=2
)
release_environment2 = ReleaseEnvironment.objects.create(
organization_id=org.id,
project_id=project2.id,
release_id=release2.id,
environment_id=environment2.id,
)
release_project_environment2 = ReleaseProjectEnvironment.objects.create(
release_id=release2.id, project_id=project2.id, environment_id=environment2.id
)
group_release2 = GroupRelease.objects.create(
project_id=project2.id, release_id=release2.id, group_id=2
)
group2 = self.create_group(project=project2, first_release=release2)
group_resolution2 = GroupResolution.objects.create(group=group2, release=release2)
# merge from #2
project3 = self.create_project(organization=org, name='baz')
environment3 = Environment.get_or_create(project=project3, name='env3')
release3 = Release.objects.create(version='cccccc', organization=org)
release3.add_project(project3)
release_commit3 = ReleaseCommit.objects.create(
organization_id=org.id, release=release2, commit=commit2, order=3
)
release_environment3 = ReleaseEnvironment.objects.create(
organization_id=org.id,
project_id=project3.id,
release_id=release3.id,
environment_id=environment3.id,
)
release_project_environment3 = ReleaseProjectEnvironment.objects.create(
release_id=release3.id, project_id=project3.id, environment_id=environment3.id
)
group_release3 = GroupRelease.objects.create(
project_id=project3.id, release_id=release3.id, group_id=3
)
group3 = self.create_group(project=project3, first_release=release3)
group_resolution3 = GroupResolution.objects.create(group=group3, release=release3)
Release.merge(release, [release2, release3])
# ReleaseCommit.release
assert ReleaseCommit.objects.get(id=release_commit.id).release == release
# should not exist because they referenced the same commit
assert not ReleaseCommit.objects.filter(id=release_commit2.id).exists()
assert ReleaseCommit.objects.get(id=release_commit3.id).release == release
# ReleaseEnvironment.release_id
assert ReleaseEnvironment.objects.get(id=release_environment.id).release_id == release.id
assert ReleaseEnvironment.objects.get(id=release_environment2.id).release_id == release.id
assert ReleaseEnvironment.objects.get(id=release_environment3.id).release_id == release.id
# ReleaseProject.release
assert release.projects.count() == 3
assert ReleaseProject.objects.filter(release=release, project=project).exists()
assert ReleaseProject.objects.filter(release=release, project=project2).exists()
assert ReleaseProject.objects.filter(release=release, project=project3).exists()
# ReleaseProjectEnvironment.release
assert ReleaseProjectEnvironment.objects.get(
id=release_project_environment.id).release_id == release.id
assert ReleaseProjectEnvironment.objects.get(
id=release_project_environment2.id).release_id == release.id
assert ReleaseProjectEnvironment.objects.get(
id=release_project_environment3.id).release_id == release.id
# GroupRelease.release_id
assert GroupRelease.objects.get(id=group_release.id).release_id == release.id
assert GroupRelease.objects.get(id=group_release2.id).release_id == release.id
assert GroupRelease.objects.get(id=group_release3.id).release_id == release.id
# GroupResolution.release
assert GroupResolution.objects.get(id=group_resolution.id).release == release
assert GroupResolution.objects.get(id=group_resolution2.id).release == release
assert GroupResolution.objects.get(id=group_resolution3.id).release == release
# Group.first_release
assert Group.objects.get(id=group.id).first_release == release
assert Group.objects.get(id=group2.id).first_release == release
assert Group.objects.get(id=group3.id).first_release == release
# Releases are gone
assert Release.objects.filter(id=release.id).exists()
assert not Release.objects.filter(id=release2.id).exists()
assert not Release.objects.filter(id=release3.id).exists()
def save(self, project, raw=False):
project = Project.objects.get_from_cache(id=project)
data = self.data.copy()
# First we pull out our top-level (non-data attr) kwargs
event_id = data.pop('event_id')
message = data.pop('message')
level = data.pop('level')
culprit = data.pop('culprit', None)
time_spent = data.pop('time_spent', None)
logger_name = data.pop('logger', None)
server_name = data.pop('server_name', None)
site = data.pop('site', None)
checksum = data.pop('checksum', None)
fingerprint = data.pop('fingerprint', None)
platform = data.pop('platform', None)
release = data.pop('release', None)
environment = data.pop('environment', None)
if not culprit:
culprit = generate_culprit(data)
date = datetime.fromtimestamp(data.pop('timestamp'))
date = date.replace(tzinfo=timezone.utc)
kwargs = {
'message': message,
'platform': platform,
}
event = Event(
project=project,
event_id=event_id,
data=data,
time_spent=time_spent,
datetime=date,
**kwargs
)
tags = data.get('tags') or []
tags.append(('level', LOG_LEVELS[level]))
if logger_name:
tags.append(('logger', logger_name))
if server_name:
tags.append(('server_name', server_name))
if site:
tags.append(('site', site))
if release:
# TODO(dcramer): we should ensure we create Release objects
tags.append(('sentry:release', release))
if environment:
tags.append(('environment', environment))
for plugin in plugins.for_project(project, version=None):
added_tags = safe_execute(plugin.get_tags, event,
_with_transaction=False)
if added_tags:
tags.extend(added_tags)
event_user = self._get_event_user(project, data)
if event_user:
tags.append(('sentry:user', event_user.tag_value))
# XXX(dcramer): we're relying on mutation of the data object to ensure
# this propagates into Event
data['tags'] = tags
data['fingerprint'] = fingerprint or '{{ default }}'
# prioritize fingerprint over checksum as its likely the client defaulted
# a checksum whereas the fingerprint was explicit
if fingerprint:
hashes = map(md5_from_hash, get_hashes_from_fingerprint(event, fingerprint))
elif checksum:
hashes = [checksum]
else:
hashes = map(md5_from_hash, get_hashes_for_event(event))
group_kwargs = kwargs.copy()
group_kwargs.update({
'culprit': culprit,
'logger': logger_name,
'level': level,
'last_seen': date,
'first_seen': date,
'time_spent_total': time_spent or 0,
'time_spent_count': time_spent and 1 or 0,
})
if release:
release = Release.get_or_create(
project=project,
version=release,
date_added=date,
)
group_kwargs['first_release'] = release
Activity.objects.create(
type=Activity.RELEASE,
project=project,
ident=release,
data={'version': release},
datetime=date,
)
group, is_new, is_regression, is_sample = self._save_aggregate(
event=event,
hashes=hashes,
release=release,
**group_kwargs
)
event.group = group
event.group_id = group.id
# store a reference to the group id to guarantee validation of isolation
event.data.bind_ref(event)
try:
with transaction.atomic():
EventMapping.objects.create(
project=project, group=group, event_id=event_id)
except IntegrityError:
self.logger.info('Duplicate EventMapping found for event_id=%s', event_id)
return event
UserReport.objects.filter(
project=project, event_id=event_id,
).update(group=group)
# save the event unless its been sampled
if not is_sample:
try:
with transaction.atomic():
event.save()
except IntegrityError:
self.logger.info('Duplicate Event found for event_id=%s', event_id)
return event
if event_user:
tsdb.record_multi((
(tsdb.models.users_affected_by_group, group.id, (event_user.tag_value,)),
(tsdb.models.users_affected_by_project, project.id, (event_user.tag_value,)),
), timestamp=event.datetime)
if is_new and release:
buffer.incr(Release, {'new_groups': 1}, {
'id': release.id,
})
safe_execute(Group.objects.add_tags, group, tags,
_with_transaction=False)
if not raw:
if not project.first_event:
project.update(first_event=date)
post_process_group.delay(
group=group,
event=event,
is_new=is_new,
is_sample=is_sample,
is_regression=is_regression,
)
else:
self.logger.info('Raw event passed; skipping post process for event_id=%s', event_id)
index_event.delay(event)
# TODO: move this to the queue
if is_regression and not raw:
regression_signal.send_robust(sender=Group, instance=group)
return event
def save(self, project, raw=False):
from sentry.tasks.post_process import index_event_tags
project = Project.objects.get_from_cache(id=project)
data = self.data.copy()
# First we pull out our top-level (non-data attr) kwargs
event_id = data.pop('event_id')
level = data.pop('level')
culprit = data.pop('culprit', None)
logger_name = data.pop('logger', None)
server_name = data.pop('server_name', None)
site = data.pop('site', None)
checksum = data.pop('checksum', None)
fingerprint = data.pop('fingerprint', None)
platform = data.pop('platform', None)
release = data.pop('release', None)
environment = data.pop('environment', None)
# unused
time_spent = data.pop('time_spent', None)
message = data.pop('message', '')
if not culprit:
culprit = generate_culprit(data, platform=platform)
date = datetime.fromtimestamp(data.pop('timestamp'))
date = date.replace(tzinfo=timezone.utc)
kwargs = {
'platform': platform,
}
event = Event(
project_id=project.id,
event_id=event_id,
data=data,
time_spent=time_spent,
datetime=date,
**kwargs
)
tags = data.get('tags') or []
tags.append(('level', LOG_LEVELS[level]))
if logger_name:
tags.append(('logger', logger_name))
if server_name:
tags.append(('server_name', server_name))
if site:
tags.append(('site', site))
if release:
# TODO(dcramer): we should ensure we create Release objects
tags.append(('sentry:release', release))
if environment:
tags.append(('environment', environment))
for plugin in plugins.for_project(project, version=None):
added_tags = safe_execute(plugin.get_tags, event,
_with_transaction=False)
if added_tags:
tags.extend(added_tags)
event_user = self._get_event_user(project, data)
if event_user:
tags.append(('sentry:user', event_user.tag_value))
# XXX(dcramer): we're relying on mutation of the data object to ensure
# this propagates into Event
data['tags'] = tags
data['fingerprint'] = fingerprint or ['{{ default }}']
for path, iface in event.interfaces.iteritems():
data['tags'].extend(iface.iter_tags())
# Get rid of ephemeral interface data
if iface.ephemeral:
data.pop(iface.get_path(), None)
# prioritize fingerprint over checksum as its likely the client defaulted
# a checksum whereas the fingerprint was explicit
if fingerprint:
hashes = map(md5_from_hash, get_hashes_from_fingerprint(event, fingerprint))
elif checksum:
hashes = [checksum]
else:
hashes = map(md5_from_hash, get_hashes_for_event(event))
# TODO(dcramer): temp workaround for complexity
data['message'] = message
event_type = eventtypes.get(data.get('type', 'default'))(data)
event_metadata = event_type.get_metadata()
# TODO(dcramer): temp workaround for complexity
del data['message']
data['type'] = event_type.key
data['metadata'] = event_metadata
# index components into ``Event.message``
# See GH-3248
if event_type.key != 'default':
if 'sentry.interfaces.Message' in data and \
data['sentry.interfaces.Message']['message'] != message:
message = u'{} {}'.format(
message,
data['sentry.interfaces.Message']['message'],
)
if not message:
message = ''
elif not isinstance(message, basestring):
message = force_text(message)
for value in event_metadata.itervalues():
value_u = force_text(value, errors='replace')
if value_u not in message:
message = u'{} {}'.format(message, value_u)
message = trim(message.strip(), settings.SENTRY_MAX_MESSAGE_LENGTH)
event.message = message
kwargs['message'] = message
group_kwargs = kwargs.copy()
group_kwargs.update({
'culprit': culprit,
'logger': logger_name,
'level': level,
'last_seen': date,
'first_seen': date,
'data': {
'last_received': event.data.get('received') or float(event.datetime.strftime('%s')),
'type': event_type.key,
# we cache the events metadata on the group to ensure its
# accessible in the stream
'metadata': event_metadata,
},
})
if release:
release = Release.get_or_create(
project=project,
version=release,
date_added=date,
)
group_kwargs['first_release'] = release
group, is_new, is_regression, is_sample = self._save_aggregate(
event=event,
hashes=hashes,
release=release,
**group_kwargs
)
event.group = group
# store a reference to the group id to guarantee validation of isolation
event.data.bind_ref(event)
try:
with transaction.atomic(using=router.db_for_write(EventMapping)):
EventMapping.objects.create(
project=project, group=group, event_id=event_id)
except IntegrityError:
self.logger.info('Duplicate EventMapping found for event_id=%s', event_id,
exc_info=True)
return event
if release:
grouprelease = GroupRelease.get_or_create(
group=group,
release=release,
environment=environment,
datetime=date,
)
tsdb.incr_multi([
(tsdb.models.group, group.id),
(tsdb.models.project, project.id),
], timestamp=event.datetime)
frequencies = [
(tsdb.models.frequent_projects_by_organization, {
project.organization_id: {
project.id: 1,
},
}),
(tsdb.models.frequent_issues_by_project, {
project.id: {
group.id: 1,
},
})
]
if release:
frequencies.append(
(tsdb.models.frequent_releases_by_groups, {
group.id: {
grouprelease.id: 1,
},
})
)
tsdb.record_frequency_multi(frequencies, timestamp=event.datetime)
UserReport.objects.filter(
project=project, event_id=event_id,
).update(group=group)
# save the event unless its been sampled
if not is_sample:
try:
with transaction.atomic(using=router.db_for_write(Event)):
event.save()
except IntegrityError:
self.logger.info('Duplicate Event found for event_id=%s', event_id,
exc_info=True)
return event
index_event_tags.delay(
project_id=project.id,
group_id=group.id,
event_id=event.id,
tags=tags,
)
if event_user:
tsdb.record_multi((
(tsdb.models.users_affected_by_group, group.id, (event_user.tag_value,)),
(tsdb.models.users_affected_by_project, project.id, (event_user.tag_value,)),
), timestamp=event.datetime)
if is_new and release:
buffer.incr(Release, {'new_groups': 1}, {
'id': release.id,
})
safe_execute(Group.objects.add_tags, group, tags,
_with_transaction=False)
if not raw:
if not project.first_event:
project.update(first_event=date)
first_event_received.send(project=project, group=group, sender=Project)
post_process_group.delay(
group=group,
event=event,
is_new=is_new,
is_sample=is_sample,
is_regression=is_regression,
)
else:
self.logger.info('Raw event passed; skipping post process for event_id=%s', event_id)
# TODO: move this to the queue
if is_regression and not raw:
regression_signal.send_robust(sender=Group, instance=group)
return event
def get_release(self, project, data):
if not data.get("release"):
return
return Release.get(project=project, version=data["release"])
def save(self, project_id, raw=False, assume_normalized=False):
# Normalize if needed
if not self._normalized:
if not assume_normalized:
self.normalize()
self._normalized = True
data = self._data
project = Project.objects.get_from_cache(id=project_id)
project._organization_cache = Organization.objects.get_from_cache(
id=project.organization_id)
# Check to make sure we're not about to do a bunch of work that's
# already been done if we've processed an event with this ID. (This
# isn't a perfect solution -- this doesn't handle ``EventMapping`` and
# there's a race condition between here and when the event is actually
# saved, but it's an improvement. See GH-7677.)
try:
event = Event.objects.get(
project_id=project.id,
event_id=data['event_id'],
)
except Event.DoesNotExist:
pass
else:
# Make sure we cache on the project before returning
event._project_cache = project
logger.info(
'duplicate.found',
exc_info=True,
extra={
'event_uuid': data['event_id'],
'project_id': project.id,
'model': Event.__name__,
}
)
return event
# Pull out the culprit
culprit = self.get_culprit()
# Pull the toplevel data we're interested in
level = data.get('level')
# TODO(mitsuhiko): this code path should be gone by July 2018.
# This is going to be fine because no code actually still depends
# on integers here. When we need an integer it will be converted
# into one later. Old workers used to send integers here.
if level is not None and isinstance(level, six.integer_types):
level = LOG_LEVELS[level]
transaction_name = data.get('transaction')
logger_name = data.get('logger')
release = data.get('release')
dist = data.get('dist')
environment = data.get('environment')
recorded_timestamp = data.get('timestamp')
# We need to swap out the data with the one internal to the newly
# created event object
event = self._get_event_instance(project_id=project_id)
self._data = data = event.data.data
event._project_cache = project
date = event.datetime
platform = event.platform
event_id = event.event_id
if transaction_name:
transaction_name = force_text(transaction_name)
# Some of the data that are toplevel attributes are duplicated
# into tags (logger, level, environment, transaction). These are
# different from legacy attributes which are normalized into tags
# ahead of time (site, server_name).
setdefault_path(data, 'tags', value=[])
set_tag(data, 'level', level)
if logger_name:
set_tag(data, 'logger', logger_name)
if environment:
set_tag(data, 'environment', environment)
if transaction_name:
set_tag(data, 'transaction', transaction_name)
if release:
# dont allow a conflicting 'release' tag
pop_tag(data, 'release')
release = Release.get_or_create(
project=project,
version=release,
date_added=date,
)
set_tag(data, 'sentry:release', release.version)
if dist and release:
dist = release.add_dist(dist, date)
# dont allow a conflicting 'dist' tag
pop_tag(data, 'dist')
set_tag(data, 'sentry:dist', dist.name)
else:
dist = None
event_user = self._get_event_user(project, data)
if event_user:
# dont allow a conflicting 'user' tag
pop_tag(data, 'user')
set_tag(data, 'sentry:user', event_user.tag_value)
# At this point we want to normalize the in_app values in case the
# clients did not set this appropriately so far.
grouping_config = load_grouping_config(
get_grouping_config_dict_for_event_data(data, project))
normalize_stacktraces_for_grouping(data, grouping_config)
for plugin in plugins.for_project(project, version=None):
added_tags = safe_execute(plugin.get_tags, event, _with_transaction=False)
if added_tags:
# plugins should not override user provided tags
for key, value in added_tags:
if get_tag(data, key) is None:
set_tag(data, key, value)
for path, iface in six.iteritems(event.interfaces):
for k, v in iface.iter_tags():
set_tag(data, k, v)
# Get rid of ephemeral interface data
if iface.ephemeral:
data.pop(iface.path, None)
# The active grouping config was put into the event in the
# normalize step before. We now also make sure that the
# fingerprint was set to `'{{ default }}' just in case someone
# removed it from the payload. The call to get_hashes will then
# look at `grouping_config` to pick the right paramters.
data['fingerprint'] = data.get('fingerprint') or ['{{ default }}']
apply_server_fingerprinting(data, get_fingerprinting_config_for_project(project))
hashes = event.get_hashes()
data['hashes'] = hashes
# we want to freeze not just the metadata and type in but also the
# derived attributes. The reason for this is that we push this
# data into kafka for snuba processing and our postprocessing
# picks up the data right from the snuba topic. For most usage
# however the data is dynamically overriden by Event.title and
# Event.location (See Event.as_dict)
materialized_metadata = self.materialize_metadata()
event_metadata = materialized_metadata['metadata']
data.update(materialized_metadata)
data['culprit'] = culprit
# index components into ``Event.message``
# See GH-3248
event.message = self.get_search_message(event_metadata, culprit)
received_timestamp = event.data.get('received') or float(event.datetime.strftime('%s'))
# The group gets the same metadata as the event when it's flushed but
# additionally the `last_received` key is set. This key is used by
# _save_aggregate.
group_metadata = dict(materialized_metadata)
group_metadata['last_received'] = received_timestamp
kwargs = {
'platform': platform,
'message': event.message,
'culprit': culprit,
'logger': logger_name,
'level': LOG_LEVELS_MAP.get(level),
'last_seen': date,
'first_seen': date,
'active_at': date,
'data': group_metadata,
}
if release:
kwargs['first_release'] = release
try:
group, is_new, is_regression, is_sample = self._save_aggregate(
event=event, hashes=hashes, release=release, **kwargs
)
except HashDiscarded:
event_discarded.send_robust(
project=project,
sender=EventManager,
)
metrics.incr(
'events.discarded',
skip_internal=True,
tags={
'organization_id': project.organization_id,
'platform': platform,
},
)
raise
else:
event_saved.send_robust(
project=project,
event_size=event.size,
sender=EventManager,
)
event.group = group
# store a reference to the group id to guarantee validation of isolation
event.data.bind_ref(event)
# When an event was sampled, the canonical source of truth
# is the EventMapping table since we aren't going to be writing out an actual
# Event row. Otherwise, if the Event isn't being sampled, we can safely
# rely on the Event table itself as the source of truth and ignore
# EventMapping since it's redundant information.
if is_sample:
try:
with transaction.atomic(using=router.db_for_write(EventMapping)):
EventMapping.objects.create(project=project, group=group, event_id=event_id)
except IntegrityError:
logger.info(
'duplicate.found',
exc_info=True,
extra={
'event_uuid': event_id,
'project_id': project.id,
'group_id': group.id,
'model': EventMapping.__name__,
}
)
return event
environment = Environment.get_or_create(
project=project,
name=environment,
)
group_environment, is_new_group_environment = GroupEnvironment.get_or_create(
group_id=group.id,
environment_id=environment.id,
defaults={
'first_release': release if release else None,
},
)
if release:
ReleaseEnvironment.get_or_create(
project=project,
release=release,
environment=environment,
datetime=date,
)
ReleaseProjectEnvironment.get_or_create(
project=project,
release=release,
environment=environment,
datetime=date,
)
grouprelease = GroupRelease.get_or_create(
group=group,
release=release,
environment=environment,
datetime=date,
)
counters = [
(tsdb.models.group, group.id),
(tsdb.models.project, project.id),
]
if release:
counters.append((tsdb.models.release, release.id))
tsdb.incr_multi(counters, timestamp=event.datetime, environment_id=environment.id)
frequencies = [
# (tsdb.models.frequent_projects_by_organization, {
# project.organization_id: {
# project.id: 1,
# },
# }),
# (tsdb.models.frequent_issues_by_project, {
# project.id: {
# group.id: 1,
# },
# })
(tsdb.models.frequent_environments_by_group, {
group.id: {
environment.id: 1,
},
})
]
if release:
frequencies.append(
(tsdb.models.frequent_releases_by_group, {
group.id: {
grouprelease.id: 1,
},
})
)
tsdb.record_frequency_multi(frequencies, timestamp=event.datetime)
UserReport.objects.filter(
project=project,
event_id=event_id,
).update(
group=group,
environment=environment,
)
# save the event unless its been sampled
if not is_sample:
try:
with transaction.atomic(using=router.db_for_write(Event)):
event.save()
except IntegrityError:
logger.info(
'duplicate.found',
exc_info=True,
extra={
'event_uuid': event_id,
'project_id': project.id,
'group_id': group.id,
'model': Event.__name__,
}
)
return event
tagstore.delay_index_event_tags(
organization_id=project.organization_id,
project_id=project.id,
group_id=group.id,
environment_id=environment.id,
event_id=event.id,
tags=event.tags,
date_added=event.datetime,
)
if event_user:
tsdb.record_multi(
(
(tsdb.models.users_affected_by_group, group.id, (event_user.tag_value, )),
(tsdb.models.users_affected_by_project, project.id, (event_user.tag_value, )),
),
timestamp=event.datetime,
environment_id=environment.id,
)
if release:
if is_new:
buffer.incr(
ReleaseProject, {'new_groups': 1}, {
'release_id': release.id,
'project_id': project.id,
}
)
if is_new_group_environment:
buffer.incr(
ReleaseProjectEnvironment, {'new_issues_count': 1}, {
'project_id': project.id,
'release_id': release.id,
'environment_id': environment.id,
}
)
safe_execute(
Group.objects.add_tags,
group,
environment,
event.get_tags(),
_with_transaction=False)
if not raw:
if not project.first_event:
project.update(first_event=date)
first_event_received.send_robust(project=project, group=group, sender=Project)
eventstream.insert(
group=group,
event=event,
is_new=is_new,
is_sample=is_sample,
is_regression=is_regression,
is_new_group_environment=is_new_group_environment,
primary_hash=hashes[0],
# We are choosing to skip consuming the event back
# in the eventstream if it's flagged as raw.
# This means that we want to publish the event
# through the event stream, but we don't care
# about post processing and handling the commit.
skip_consume=raw,
)
metrics.timing(
'events.latency',
received_timestamp - recorded_timestamp,
tags={
'project_id': project.id,
},
)
metrics.timing(
'events.size.data.post_save',
event.size,
tags={'project_id': project.id}
)
return event
def validate_version(self, value):
if not Release.is_valid_version(value):
raise serializers.ValidationError("Release with name %s is not allowed" % value)
return value
def save(self, project, raw=False):
# TODO: culprit should default to "most recent" frame in stacktraces when
# it's not provided.
project = Project.objects.get_from_cache(id=project)
data = self.data.copy()
# First we pull out our top-level (non-data attr) kwargs
event_id = data.pop('event_id')
message = data.pop('message')
level = data.pop('level')
culprit = data.pop('culprit', None) or ''
time_spent = data.pop('time_spent', None)
logger_name = data.pop('logger', None)
server_name = data.pop('server_name', None)
site = data.pop('site', None)
checksum = data.pop('checksum', None)
platform = data.pop('platform', None)
release = data.pop('release', None)
date = datetime.fromtimestamp(data.pop('timestamp'))
date = date.replace(tzinfo=timezone.utc)
kwargs = {
'message': message,
'platform': platform,
}
event = Event(
project=project,
event_id=event_id,
data=data,
time_spent=time_spent,
datetime=date,
**kwargs
)
tags = data.get('tags') or []
tags.append(('level', LOG_LEVELS[level]))
if logger_name:
tags.append(('logger', logger_name))
if server_name:
tags.append(('server_name', server_name))
if site:
tags.append(('site', site))
if release:
# TODO(dcramer): we should ensure we create Release objects
tags.append(('sentry:release', release))
for plugin in plugins.for_project(project, version=None):
added_tags = safe_execute(plugin.get_tags, event,
_with_transaction=False)
if added_tags:
tags.extend(added_tags)
# XXX(dcramer): we're relying on mutation of the data object to ensure
# this propagates into Event
data['tags'] = tags
# Calculate the checksum from the first highest scoring interface
if checksum:
hashes = [checksum]
else:
hashes = get_hashes_for_event(event)
group_kwargs = kwargs.copy()
group_kwargs.update({
'culprit': culprit,
'logger': logger_name,
'level': level,
'last_seen': date,
'first_seen': date,
'time_spent_total': time_spent or 0,
'time_spent_count': time_spent and 1 or 0,
})
if release:
group_kwargs['first_release'] = Release.get_or_create(
project=project,
version=release,
date_added=date,
)
Activity.objects.create(
type=Activity.RELEASE,
project=project,
ident=release,
data={'version': release},
datetime=date,
)
group, is_new, is_regression, is_sample = safe_execute(
self._save_aggregate,
event=event,
hashes=hashes,
**group_kwargs
)
using = group._state.db
event.group = group
try:
with transaction.atomic():
EventMapping.objects.create(
project=project, group=group, event_id=event_id)
except IntegrityError:
self.logger.info('Duplicate EventMapping found for event_id=%s', event_id)
return event
UserReport.objects.filter(
project=project, event_id=event_id,
).update(group=group)
# save the event unless its been sampled
if not is_sample:
try:
with transaction.atomic():
event.save()
except IntegrityError:
self.logger.info('Duplicate Event found for event_id=%s', event_id)
return event
safe_execute(Group.objects.add_tags, group, tags,
_with_transaction=False)
if not raw:
post_process_group.delay(
group=group,
event=event,
is_new=is_new,
is_sample=is_sample,
is_regression=is_regression,
)
else:
self.logger.info('Raw event passed; skipping post process for event_id=%s', event_id)
index_event.delay(event)
# TODO: move this to the queue
if is_regression and not raw:
regression_signal.send_robust(sender=Group, instance=group)
return event
def debounce_update_release_health_data(organization, project_ids):
"""This causes a flush of snuba health data to the postgres tables once
per minute for the given projects.
"""
# Figure out which projects need to get updates from the snuba.
should_update = {}
cache_keys = ["debounce-health:%d" % id for id in project_ids]
cache_data = cache.get_many(cache_keys)
for project_id, cache_key in izip(project_ids, cache_keys):
if cache_data.get(cache_key) is None:
should_update[project_id] = cache_key
if not should_update:
return
projects = {
p.id: p
for p in Project.objects.get_many_from_cache(should_update.keys())
}
# This gives us updates for all release-projects which have seen new
# health data over the last days. It will miss releases where the last
# date is longer than what `get_changed_project_release_model_adoptions`
# considers recent.
project_releases = release_health.get_changed_project_release_model_adoptions(
should_update.keys())
# Check which we already have rows for.
existing = set(
ReleaseProject.objects.filter(
project_id__in=[x[0] for x in project_releases],
release__version__in=[x[1] for x in project_releases],
).values_list("project_id", "release__version"))
to_upsert = []
for key in project_releases:
if key not in existing:
to_upsert.append(key)
if to_upsert:
dates = release_health.get_oldest_health_data_for_releases(to_upsert)
for project_id, version in to_upsert:
project = projects.get(project_id)
if project is None:
# should not happen
continue
# We might have never observed the release. This for instance can
# happen if the release only had health data so far. For these cases
# we want to create the release the first time we observed it on the
# health side.
release = Release.get_or_create(project=project,
version=version,
date_added=dates.get(
(project_id, version)))
# Make sure that the release knows about this project. Like we had before
# the project might not have been associated with this release yet.
release.add_project(project)
# Debounce updates for a minute
cache.set_many(
dict(izip(should_update.values(), [True] * len(should_update))), 60)
def create_message_event(template, parameters, environment, release):
i = next(sequence)
event_id = uuid.UUID(fields=(
i,
0x0,
0x1000,
0x80,
0x80,
0x808080808080,
), ).hex
tags = [['color', next(tag_values)]]
if environment:
tags.append(['environment', environment])
if release:
tags.append(['sentry:release', release])
event = Event.objects.create(
project_id=project.id,
group_id=source.id,
event_id=event_id,
message='%s' % (id, ),
datetime=now + shift(i),
data={
'environment': environment,
'type': 'default',
'metadata': {
'title': template % parameters,
},
'logentry': {
'message': template,
'params': parameters,
'formatted': template % parameters,
},
'user': next(user_values),
'tags': tags,
},
)
with self.tasks():
Group.objects.add_tags(
source,
Environment.objects.get(
organization_id=project.organization_id,
name=environment),
tags=event.tags,
)
EventMapping.objects.create(
project_id=project.id,
group_id=source.id,
event_id=event_id,
date_added=event.datetime,
)
UserReport.objects.create(
project_id=project.id,
group_id=source.id,
event_id=event_id,
name='Log Hat',
email='*****@*****.**',
comments='Quack',
)
if release:
Release.get_or_create(
project=project,
version=event.get_tag('sentry:release'),
date_added=event.datetime,
)
features.record([event])
return event
def test_simple(self):
org = self.create_organization()
commit = Commit.objects.create(organization_id=org.id, repository_id=5)
commit2 = Commit.objects.create(organization_id=org.id,
repository_id=6)
# merge to
project = self.create_project(organization=org, name='foo')
release = Release.objects.create(version='abcdabc', organization=org)
release.add_project(project)
release_commit = ReleaseCommit.objects.create(organization_id=org.id,
release=release,
commit=commit,
order=1)
release_environment = ReleaseEnvironment.objects.create(
organization_id=org.id,
project_id=project.id,
release_id=release.id,
environment_id=2)
group_release = GroupRelease.objects.create(project_id=project.id,
release_id=release.id,
group_id=1)
group = self.create_group(project=project, first_release=release)
group_resolution = GroupResolution.objects.create(group=group,
release=release)
# merge from #1
project2 = self.create_project(organization=org, name='bar')
release2 = Release.objects.create(version='bbbbbbb', organization=org)
release2.add_project(project2)
release_commit2 = ReleaseCommit.objects.create(organization_id=org.id,
release=release2,
commit=commit,
order=2)
release_environment2 = ReleaseEnvironment.objects.create(
organization_id=org.id,
project_id=project2.id,
release_id=release2.id,
environment_id=3,
)
group_release2 = GroupRelease.objects.create(project_id=project2.id,
release_id=release2.id,
group_id=2)
group2 = self.create_group(project=project2, first_release=release2)
group_resolution2 = GroupResolution.objects.create(group=group2,
release=release2)
# merge from #2
project3 = self.create_project(organization=org, name='baz')
release3 = Release.objects.create(version='cccccc', organization=org)
release3.add_project(project3)
release_commit3 = ReleaseCommit.objects.create(organization_id=org.id,
release=release2,
commit=commit2,
order=3)
release_environment3 = ReleaseEnvironment.objects.create(
organization_id=org.id,
project_id=project3.id,
release_id=release3.id,
environment_id=4,
)
group_release3 = GroupRelease.objects.create(project_id=project3.id,
release_id=release3.id,
group_id=3)
group3 = self.create_group(project=project3, first_release=release3)
group_resolution3 = GroupResolution.objects.create(group=group3,
release=release3)
Release.merge(release, [release2, release3])
# ReleaseCommit.release
assert ReleaseCommit.objects.get(
id=release_commit.id).release == release
# should not exist because they referenced the same commit
assert not ReleaseCommit.objects.filter(id=release_commit2.id).exists()
assert ReleaseCommit.objects.get(
id=release_commit3.id).release == release
# ReleaseEnvironment.release_id
assert ReleaseEnvironment.objects.get(
id=release_environment.id).release_id == release.id
assert ReleaseEnvironment.objects.get(
id=release_environment2.id).release_id == release.id
assert ReleaseEnvironment.objects.get(
id=release_environment3.id).release_id == release.id
# ReleaseProject.release
assert release.projects.count() == 3
assert ReleaseProject.objects.filter(release=release,
project=project).exists()
assert ReleaseProject.objects.filter(release=release,
project=project2).exists()
assert ReleaseProject.objects.filter(release=release,
project=project3).exists()
# GroupRelease.release_id
assert GroupRelease.objects.get(
id=group_release.id).release_id == release.id
assert GroupRelease.objects.get(
id=group_release2.id).release_id == release.id
assert GroupRelease.objects.get(
id=group_release3.id).release_id == release.id
# GroupResolution.release
assert GroupResolution.objects.get(
id=group_resolution.id).release == release
assert GroupResolution.objects.get(
id=group_resolution2.id).release == release
assert GroupResolution.objects.get(
id=group_resolution3.id).release == release
# Group.first_release
assert Group.objects.get(id=group.id).first_release == release
assert Group.objects.get(id=group2.id).first_release == release
assert Group.objects.get(id=group3.id).first_release == release
# Releases are gone
assert Release.objects.filter(id=release.id).exists()
assert not Release.objects.filter(id=release2.id).exists()
assert not Release.objects.filter(id=release3.id).exists()
def get(self, request):
org = Organization(
id=1,
slug='organization',
name='My Company',
)
projects = [
Project(
id=1,
organization=org,
slug='project',
name='My Project',
),
Project(
id=2,
organization=org,
slug='another-project',
name='Another Project',
),
Project(
id=3,
organization=org,
slug='yet-another-project',
name='Yet Another Project',
),
]
release = Release(organization_id=org.id,
version='6c998f755f304593a4713abd123eaf8833a2de5e',
date_added=datetime.datetime(2016,
10,
12,
15,
39,
tzinfo=pytz.utc))
deploy = Deploy(
release=release,
organization_id=org.id,
environment_id=1,
date_finished=datetime.datetime(2016,
10,
12,
15,
39,
tzinfo=pytz.utc),
)
has_new_links = features.has('organizations:sentry10', org)
if has_new_links:
release_links = [
absolute_uri(
u'/organizations/{}/releases/{}/?project={}'.format(
org.slug,
release.version,
p.id,
)) for p in projects
]
else:
release_links = [
absolute_uri(u'/{}/{}/releases/{}/'.format(
org.slug,
p.slug,
release.version,
)) for p in projects
]
repos = [{
'name':
'getsentry/getsentry',
'commits': [
(Commit(
key='48b86fcd677da3dba5679d7a738240ce6fb74b20',
date_added=datetime.datetime(2016,
10,
11,
15,
39,
tzinfo=pytz.utc),
), None),
(Commit(
key='a53a2756bb8d111b43196210b34df90b87ed336b',
message='Fix billing',
author=CommitAuthor(
name='David Cramer',
email='*****@*****.**',
),
date_added=datetime.datetime(2016,
10,
11,
16,
45,
tzinfo=pytz.utc),
), User(email='*****@*****.**', name='David Cramer')),
],
}, {
'name':
'getsentry/sentry',
'commits': [
(Commit(
key='3c8eb3b4af6ee2a29c68daa188fc730c8e4b39fd',
date_added=datetime.datetime(2016,
10,
10,
15,
39,
tzinfo=pytz.utc),
), None),
(Commit(
key='373562702009df1692da6eb80a933139f29e094b',
message='Fix padding',
author=CommitAuthor(
name='Chris Jennings',
email='*****@*****.**',
),
date_added=datetime.datetime(2016,
10,
10,
16,
39,
tzinfo=pytz.utc),
), None),
(Commit(
key='631cd9096bd9811a046a472bb0aa8b573e86e1f1',
message='Update README.rst',
author=CommitAuthor(
name='David Cramer',
email='*****@*****.**',
),
date_added=datetime.datetime(2016,
10,
11,
10,
39,
tzinfo=pytz.utc),
), User(email='*****@*****.**', name='David Cramer')),
],
}]
return MailPreview(
html_template='sentry/emails/activity/release.html',
text_template='sentry/emails/activity/release.txt',
context={
'release':
release,
'projects':
zip(projects, release_links, [6, 1, 0]),
'repos':
repos,
'reason':
GroupSubscriptionReason.descriptions[
GroupSubscriptionReason.committed],
'project_count':
len(projects),
'commit_count':
4,
'author_count':
1,
'file_count':
5,
'environment':
'production',
'deploy':
deploy,
'setup_repo_link':
absolute_uri('/organizations/{}/repos/'.format(org.slug, )),
},
).render(request)
def validate_version(self, attrs, source):
value = attrs[source]
if not Release.is_valid_version(value):
raise serializers.ValidationError('Invalid value for release')
return attrs
def merge_to(from_org, to_org):
from sentry.models import (
ApiKey,
AuditLogEntry,
AuthProvider,
Commit,
Environment,
OrganizationAvatar,
OrganizationIntegration,
OrganizationMember,
OrganizationMemberTeam,
Project,
Release,
ReleaseCommit,
ReleaseEnvironment,
ReleaseFile,
ReleaseHeadCommit,
Repository,
Team,
)
for from_member in OrganizationMember.objects.filter(
organization=from_org, user__isnull=False):
logger = logging.getLogger("sentry.merge")
try:
to_member = OrganizationMember.objects.get(
organization=to_org, user=from_member.user)
except OrganizationMember.DoesNotExist:
from_member.update(organization=to_org)
to_member = from_member
else:
qs = OrganizationMemberTeam.objects.filter(
organizationmember=from_member,
is_active=True).select_related()
for omt in qs:
OrganizationMemberTeam.objects.create_or_update(
organizationmember=to_member,
team=omt.team,
defaults={"is_active": True})
logger.info(
"user.migrate",
extra={
"instance_id": from_member.id,
"new_member_id": to_member.id,
"from_organization_id": from_org.id,
"to_organization_id": to_org.id,
},
)
for from_team in Team.objects.filter(organization=from_org):
try:
with transaction.atomic():
from_team.update(organization=to_org)
except IntegrityError:
slugify_instance(from_team,
from_team.name,
organization=to_org)
from_team.update(organization=to_org, slug=from_team.slug)
logger.info(
"team.migrate",
extra={
"instance_id": from_team.id,
"new_slug": from_team.slug,
"from_organization_id": from_org.id,
"to_organization_id": to_org.id,
},
)
for from_project in Project.objects.filter(organization=from_org):
try:
with transaction.atomic():
from_project.update(organization=to_org)
except IntegrityError:
slugify_instance(
from_project,
from_project.name,
organization=to_org,
reserved=RESERVED_PROJECT_SLUGS,
)
from_project.update(organization=to_org,
slug=from_project.slug)
logger.info(
"project.migrate",
extra={
"instance_id": from_project.id,
"new_slug": from_project.slug,
"from_organization_id": from_org.id,
"to_organization_id": to_org.id,
},
)
# TODO(jess): update this when adding unique constraint
# on version, organization for releases
for from_release in Release.objects.filter(organization=from_org):
try:
to_release = Release.objects.get(version=from_release.version,
organization=to_org)
except Release.DoesNotExist:
Release.objects.filter(id=from_release.id).update(
organization=to_org)
else:
Release.merge(to_release, [from_release])
logger.info(
"release.migrate",
extra={
"instance_id": from_release.id,
"from_organization_id": from_org.id,
"to_organization_id": to_org.id,
},
)
def do_update(queryset, params):
model_name = queryset.model.__name__.lower()
try:
with transaction.atomic():
queryset.update(**params)
except IntegrityError:
for instance in queryset:
try:
with transaction.atomic():
instance.update(**params)
except IntegrityError:
logger.info(
f"{model_name}.migrate-skipped",
extra={
"from_organization_id": from_org.id,
"to_organization_id": to_org.id,
},
)
else:
logger.info(
f"{model_name}.migrate",
extra={
"instance_id": instance.id,
"from_organization_id": from_org.id,
"to_organization_id": to_org.id,
},
)
else:
logger.info(
f"{model_name}.migrate",
extra={
"from_organization_id": from_org.id,
"to_organization_id": to_org.id
},
)
INST_MODEL_LIST = (
AuthProvider,
ApiKey,
AuditLogEntry,
OrganizationAvatar,
OrganizationIntegration,
ReleaseEnvironment,
ReleaseFile,
)
ATTR_MODEL_LIST = (Commit, ReleaseCommit, ReleaseHeadCommit,
Repository, Environment)
for model in INST_MODEL_LIST:
queryset = model.objects.filter(organization=from_org)
do_update(queryset, {"organization": to_org})
for model in ATTR_MODEL_LIST:
queryset = model.objects.filter(organization_id=from_org.id)
do_update(queryset, {"organization_id": to_org.id})
def set_commits(self, version, commit_list):
"""
Commits should be ordered oldest to newest.
Calling this method will remove all existing commit history.
"""
project = self.project
release = Release.objects.filter(
organization_id=project.organization_id,
version=version,
projects=self.project
).first()
if not release:
release = Release.objects.filter(
organization_id=project.organization_id,
version=version,
).first()
if not release:
lock_key = Release.get_lock_key(project.organization_id, version)
lock = locks.get(lock_key, duration=5)
with TimedRetryPolicy(10)(lock.acquire):
try:
release = Release.objects.get(
organization_id=project.organization_id,
version=version
)
except Release.DoesNotExist:
release = Release.objects.create(
organization_id=project.organization_id,
version=version
)
release.add_project(project)
with transaction.atomic():
# TODO(dcramer): would be good to optimize the logic to avoid these
# deletes but not overly important
ReleaseCommit.objects.filter(
release=release,
).delete()
authors = {}
repos = {}
for idx, data in enumerate(commit_list):
repo_name = data.get('repository') or 'project-{}'.format(project.id)
if repo_name not in repos:
repos[repo_name] = repo = Repository.objects.get_or_create(
organization_id=project.organization_id,
name=repo_name,
)[0]
else:
repo = repos[repo_name]
author_email = data.get('author_email')
if author_email is None and data.get('author_name'):
author_email = self._to_email(data['author_name'])
if not author_email:
author = None
elif author_email not in authors:
authors[author_email] = author = CommitAuthor.objects.get_or_create(
organization_id=project.organization_id,
email=author_email,
defaults={
'name': data.get('author_name'),
}
)[0]
if data.get('author_name') and author.name != data['author_name']:
author.update(name=data['author_name'])
else:
author = authors[author_email]
commit = Commit.objects.get_or_create(
organization_id=project.organization_id,
repository_id=repo.id,
key=data['id'],
defaults={
'message': data.get('message'),
'author': author,
'date_added': data.get('timestamp') or timezone.now(),
}
)[0]
ReleaseCommit.objects.create(
organization_id=project.organization_id,
release=release,
commit=commit,
order=idx,
)
def post(self, request, project):
"""
Create a New Release
````````````````````
Create a new release and/or associate a project with a release.
Release versions that are the same across multiple projects
within an Organization will be treated as the same release in Sentry.
Releases are used by Sentry to improve its error reporting abilities
by correlating first seen events with the release that might have
introduced the problem.
Releases are also necessary for sourcemaps and other debug features
that require manual upload for functioning well.
:pparam string organization_slug: the slug of the organization the
release belongs to.
:pparam string project_slug: the slug of the project to create a
release for.
:param string version: a version identifier for this release. Can
be a version number, a commit hash etc.
:param string ref: an optional commit reference. This is useful if
a tagged version has been provided.
:param url url: a URL that points to the release. This can be the
path to an online interface to the sourcecode
for instance.
:param datetime dateStarted: an optional date that indicates when the
release process started.
:param datetime dateReleased: an optional date that indicates when
the release went live. If not provided
the current time is assumed.
:auth: required
"""
serializer = ReleaseSerializer(data=request.DATA)
if serializer.is_valid():
result = serializer.object
# release creation is idempotent to simplify user
# experiences
release = Release.objects.filter(
organization_id=project.organization_id,
version=result['version'],
projects=project
).first()
created = False
if release:
was_released = bool(release.date_released)
else:
release = Release.objects.filter(
organization_id=project.organization_id,
version=result['version'],
).first()
if not release:
lock_key = Release.get_lock_key(project.organization_id, result['version'])
lock = locks.get(lock_key, duration=5)
with TimedRetryPolicy(10)(lock.acquire):
try:
release, created = Release.objects.get(
version=result['version'],
organization_id=project.organization_id
), False
except Release.DoesNotExist:
release, created = Release.objects.create(
organization_id=project.organization_id,
version=result['version'],
ref=result.get('ref'),
url=result.get('url'),
owner=result.get('owner'),
date_started=result.get('dateStarted'),
date_released=result.get('dateReleased'),
), True
was_released = False
try:
with transaction.atomic():
ReleaseProject.objects.create(project=project, release=release)
created = True
except IntegrityError:
pass
commit_list = result.get('commits')
if commit_list:
hook = ReleaseHook(project)
# TODO(dcramer): handle errors with release payloads
hook.set_commits(release.version, commit_list)
if (not was_released and release.date_released):
activity = Activity.objects.create(
type=Activity.RELEASE,
project=project,
ident=result['version'],
data={'version': result['version']},
datetime=release.date_released,
)
activity.send_notification()
if not created:
# This is the closest status code that makes sense, and we want
# a unique 2xx response code so people can understand when
# behavior differs.
# 208 Already Reported (WebDAV; RFC 5842)
status = 208
else:
status = 201
return Response(serialize(release, request.user), status=status)
return Response(serializer.errors, status=400)
def save(self, project_id, raw=False, assume_normalized=False):
# Normalize if needed
if not self._normalized:
if not assume_normalized:
self.normalize()
self._normalized = True
data = self._data
project = Project.objects.get_from_cache(id=project_id)
project._organization_cache = Organization.objects.get_from_cache(
id=project.organization_id)
# Check to make sure we're not about to do a bunch of work that's
# already been done if we've processed an event with this ID. (This
# isn't a perfect solution -- this doesn't handle ``EventMapping`` and
# there's a race condition between here and when the event is actually
# saved, but it's an improvement. See GH-7677.)
try:
event = Event.objects.get(
project_id=project.id,
event_id=data['event_id'],
)
except Event.DoesNotExist:
pass
else:
# Make sure we cache on the project before returning
event._project_cache = project
logger.info(
'duplicate.found',
exc_info=True,
extra={
'event_uuid': data['event_id'],
'project_id': project.id,
'model': Event.__name__,
}
)
return event
# Pull out the culprit
culprit = self.get_culprit()
# Pull the toplevel data we're interested in
level = data.get('level')
# TODO(mitsuhiko): this code path should be gone by July 2018.
# This is going to be fine because no code actually still depends
# on integers here. When we need an integer it will be converted
# into one later. Old workers used to send integers here.
if level is not None and isinstance(level, six.integer_types):
level = LOG_LEVELS[level]
transaction_name = data.get('transaction')
logger_name = data.get('logger')
release = data.get('release')
dist = data.get('dist')
environment = data.get('environment')
recorded_timestamp = data.get('timestamp')
# We need to swap out the data with the one internal to the newly
# created event object
event = self._get_event_instance(project_id=project_id)
self._data = data = event.data.data
event._project_cache = project
date = event.datetime
platform = event.platform
event_id = event.event_id
if transaction_name:
transaction_name = force_text(transaction_name)
# Some of the data that are toplevel attributes are duplicated
# into tags (logger, level, environment, transaction). These are
# different from legacy attributes which are normalized into tags
# ahead of time (site, server_name).
setdefault_path(data, 'tags', value=[])
set_tag(data, 'level', level)
if logger_name:
set_tag(data, 'logger', logger_name)
if environment:
set_tag(data, 'environment', environment)
if transaction_name:
set_tag(data, 'transaction', transaction_name)
if release:
# dont allow a conflicting 'release' tag
pop_tag(data, 'release')
release = Release.get_or_create(
project=project,
version=release,
date_added=date,
)
set_tag(data, 'sentry:release', release.version)
if dist and release:
dist = release.add_dist(dist, date)
# dont allow a conflicting 'dist' tag
pop_tag(data, 'dist')
set_tag(data, 'sentry:dist', dist.name)
else:
dist = None
event_user = self._get_event_user(project, data)
if event_user:
# dont allow a conflicting 'user' tag
pop_tag(data, 'user')
set_tag(data, 'sentry:user', event_user.tag_value)
# At this point we want to normalize the in_app values in case the
# clients did not set this appropriately so far.
grouping_config = load_grouping_config(
get_grouping_config_dict_for_event_data(data, project))
normalize_stacktraces_for_grouping(data, grouping_config)
for plugin in plugins.for_project(project, version=None):
added_tags = safe_execute(plugin.get_tags, event, _with_transaction=False)
if added_tags:
# plugins should not override user provided tags
for key, value in added_tags:
if get_tag(data, key) is None:
set_tag(data, key, value)
for path, iface in six.iteritems(event.interfaces):
for k, v in iface.iter_tags():
set_tag(data, k, v)
# Get rid of ephemeral interface data
if iface.ephemeral:
data.pop(iface.path, None)
# The active grouping config was put into the event in the
# normalize step before. We now also make sure that the
# fingerprint was set to `'{{ default }}' just in case someone
# removed it from the payload. The call to get_hashes will then
# look at `grouping_config` to pick the right paramters.
data['fingerprint'] = data.get('fingerprint') or ['{{ default }}']
apply_server_fingerprinting(data, get_fingerprinting_config_for_project(project))
# Here we try to use the grouping config that was requested in the
# event. If that config has since been deleted (because it was an
# experimental grouping config) we fall back to the default.
try:
hashes = event.get_hashes()
except GroupingConfigNotFound:
data['grouping_config'] = get_grouping_config_dict_for_project(project)
hashes = event.get_hashes()
data['hashes'] = hashes
# we want to freeze not just the metadata and type in but also the
# derived attributes. The reason for this is that we push this
# data into kafka for snuba processing and our postprocessing
# picks up the data right from the snuba topic. For most usage
# however the data is dynamically overriden by Event.title and
# Event.location (See Event.as_dict)
materialized_metadata = self.materialize_metadata()
event_metadata = materialized_metadata['metadata']
data.update(materialized_metadata)
data['culprit'] = culprit
# index components into ``Event.message``
# See GH-3248
event.message = self.get_search_message(event_metadata, culprit)
received_timestamp = event.data.get('received') or float(event.datetime.strftime('%s'))
# The group gets the same metadata as the event when it's flushed but
# additionally the `last_received` key is set. This key is used by
# _save_aggregate.
group_metadata = dict(materialized_metadata)
group_metadata['last_received'] = received_timestamp
kwargs = {
'platform': platform,
'message': event.message,
'culprit': culprit,
'logger': logger_name,
'level': LOG_LEVELS_MAP.get(level),
'last_seen': date,
'first_seen': date,
'active_at': date,
'data': group_metadata,
}
if release:
kwargs['first_release'] = release
try:
group, is_new, is_regression, is_sample = self._save_aggregate(
event=event, hashes=hashes, release=release, **kwargs
)
except HashDiscarded:
event_discarded.send_robust(
project=project,
sender=EventManager,
)
metrics.incr(
'events.discarded',
skip_internal=True,
tags={
'organization_id': project.organization_id,
'platform': platform,
},
)
raise
else:
event_saved.send_robust(
project=project,
event_size=event.size,
sender=EventManager,
)
event.group = group
# store a reference to the group id to guarantee validation of isolation
event.data.bind_ref(event)
# When an event was sampled, the canonical source of truth
# is the EventMapping table since we aren't going to be writing out an actual
# Event row. Otherwise, if the Event isn't being sampled, we can safely
# rely on the Event table itself as the source of truth and ignore
# EventMapping since it's redundant information.
if is_sample:
try:
with transaction.atomic(using=router.db_for_write(EventMapping)):
EventMapping.objects.create(project=project, group=group, event_id=event_id)
except IntegrityError:
logger.info(
'duplicate.found',
exc_info=True,
extra={
'event_uuid': event_id,
'project_id': project.id,
'group_id': group.id,
'model': EventMapping.__name__,
}
)
return event
environment = Environment.get_or_create(
project=project,
name=environment,
)
group_environment, is_new_group_environment = GroupEnvironment.get_or_create(
group_id=group.id,
environment_id=environment.id,
defaults={
'first_release': release if release else None,
},
)
if release:
ReleaseEnvironment.get_or_create(
project=project,
release=release,
environment=environment,
datetime=date,
)
ReleaseProjectEnvironment.get_or_create(
project=project,
release=release,
environment=environment,
datetime=date,
)
grouprelease = GroupRelease.get_or_create(
group=group,
release=release,
environment=environment,
datetime=date,
)
counters = [
(tsdb.models.group, group.id),
(tsdb.models.project, project.id),
]
if release:
counters.append((tsdb.models.release, release.id))
tsdb.incr_multi(counters, timestamp=event.datetime, environment_id=environment.id)
frequencies = [
# (tsdb.models.frequent_projects_by_organization, {
# project.organization_id: {
# project.id: 1,
# },
# }),
# (tsdb.models.frequent_issues_by_project, {
# project.id: {
# group.id: 1,
# },
# })
(tsdb.models.frequent_environments_by_group, {
group.id: {
environment.id: 1,
},
})
]
if release:
frequencies.append(
(tsdb.models.frequent_releases_by_group, {
group.id: {
grouprelease.id: 1,
},
})
)
tsdb.record_frequency_multi(frequencies, timestamp=event.datetime)
UserReport.objects.filter(
project=project,
event_id=event_id,
).update(
group=group,
environment=environment,
)
# Update any event attachment that arrived before the event group was defined.
EventAttachment.objects.filter(
project_id=project.id,
event_id=event_id,
).update(
group_id=group.id,
)
# save the event unless its been sampled
if not is_sample:
try:
with transaction.atomic(using=router.db_for_write(Event)):
event.save()
except IntegrityError:
logger.info(
'duplicate.found',
exc_info=True,
extra={
'event_uuid': event_id,
'project_id': project.id,
'group_id': group.id,
'model': Event.__name__,
}
)
return event
tagstore.delay_index_event_tags(
organization_id=project.organization_id,
project_id=project.id,
group_id=group.id,
environment_id=environment.id,
event_id=event.id,
tags=event.tags,
date_added=event.datetime,
)
if event_user:
tsdb.record_multi(
(
(tsdb.models.users_affected_by_group, group.id, (event_user.tag_value, )),
(tsdb.models.users_affected_by_project, project.id, (event_user.tag_value, )),
),
timestamp=event.datetime,
environment_id=environment.id,
)
if release:
if is_new:
buffer.incr(
ReleaseProject, {'new_groups': 1}, {
'release_id': release.id,
'project_id': project.id,
}
)
if is_new_group_environment:
buffer.incr(
ReleaseProjectEnvironment, {'new_issues_count': 1}, {
'project_id': project.id,
'release_id': release.id,
'environment_id': environment.id,
}
)
safe_execute(
Group.objects.add_tags,
group,
environment,
event.get_tags(),
_with_transaction=False)
if not raw:
if not project.first_event:
project.update(first_event=date)
first_event_received.send_robust(project=project, group=group, sender=Project)
eventstream.insert(
group=group,
event=event,
is_new=is_new,
is_sample=is_sample,
is_regression=is_regression,
is_new_group_environment=is_new_group_environment,
primary_hash=hashes[0],
# We are choosing to skip consuming the event back
# in the eventstream if it's flagged as raw.
# This means that we want to publish the event
# through the event stream, but we don't care
# about post processing and handling the commit.
skip_consume=raw,
)
metrics.timing(
'events.latency',
received_timestamp - recorded_timestamp,
tags={
'project_id': project.id,
},
)
metrics.timing(
'events.size.data.post_save',
event.size,
tags={'project_id': project.id}
)
return event
def get(self, _, project, event_id):
"""
Retrieve Committer information for an event
```````````````````````````````````````````
Return commiters on an individual event, plus a per-frame breakdown.
:pparam string project_slug: the slug of the project the event
belongs to.
:pparam string event_id: the hexadecimal ID of the event to
retrieve (as reported by the raven client).
:auth: required
"""
try:
event = Event.objects.get(
id=event_id,
project_id=project.id,
)
except Event.DoesNotExist:
return Response({'detail': 'Event not found'}, status=404)
# populate event data
Event.objects.bind_nodes([event], 'data')
group = Group.objects.get(id=event.group_id)
first_release_version = group.get_first_release()
if not first_release_version:
return Response({'detail': 'Release not found'}, status=404)
releases = Release.get_closest_releases(project, first_release_version)
if not releases:
return Response({'detail': 'Release not found'}, status=404)
commits = self._get_commits(releases)
if not commits:
return Response({'detail': 'No Commits found for Release'}, status=404)
frames = self._get_frame_paths(event)
frame_limit = 15
app_frames = [frame for frame in frames if frame['in_app']][:frame_limit]
# TODO(maxbittker) return this set instead of annotated frames
path_set = {frame['abs_path'] for frame in app_frames}
file_changes = []
if path_set:
file_changes = self._get_commit_file_changes(commits, path_set)
commit_path_matches = {
path: self._match_commits_path(file_changes, path) for path in path_set
}
annotated_frames = [
{
'frame': frame,
'commits': commit_path_matches[frame['abs_path']]
} for frame in app_frames
]
relevant_commits = list(
{commit for match in commit_path_matches for commit in commit_path_matches[match]}
)
committers = self._get_committers(annotated_frames, relevant_commits)
# serialize the commit objects
serialized_annotated_frames = [
{
'frame': frame['frame'],
'commits': serialize(frame['commits'])
} for frame in annotated_frames
]
data = {
# map author ids to sentry user dicts
'committers': committers,
'annotatedFrames': serialized_annotated_frames
}
return Response(data)
def validate_version(self, attrs, source):
value = attrs[source]
if not Release.is_valid_version(value):
raise serializers.ValidationError(
'Release with name %s is not allowed' % value)
return attrs
def validate_version(self, attrs, source):
value = attrs[source]
if not Release.is_valid_version(value):
raise serializers.ValidationError('Release with name %s is not allowed' % value)
return attrs
