def test_chunks(self):
self.assertEqual(
list(utils.chunks(range(100), size=50)),
[list(range(50)), list(range(50, 100, 1))],
)
self.assertEqual(list(utils.chunks(range(1), size=50)), [[0]])
self.assertEqual(
list(utils.chunks(range(60), size=50)),
[list(range(50)), list(range(50, 60, 1))],
)
def process(self, param_groups):
client = local_session(self.manager.session_factory).client('rds')
params = []
for param in self.data.get('params', []):
params.append({
'ParameterName': param['name'],
'ParameterValue': param['value'],
'ApplyMethod': param.get('apply-method', 'immediate'),
})
for param_group in param_groups:
name = self.get_pg_name(param_group)
# Fetch the existing parameters for this DB, so we only try to change the ones that are
# different.
cur_params = self.get_current_params(client, name)
changed_params = []
for param in params:
param_name = param['ParameterName']
if (param_name not in cur_params or
cur_params[param_name]['ParameterValue'] != param['ParameterValue']):
changed_params.append(param)
# Can only do 20 elements at a time per docs, so if we have more than that we will
# break it into multiple requests: https://goo.gl/Z6oGNv
for param_set in chunks(changed_params, 5):
self.do_modify(client, name, param_set)
self.log.info('Modified RDS parameter group %s (%i parameters changed, %i unchanged)',
name, len(changed_params), len(params) - len(changed_params))
def flush(self):
buf = self.buf
self.buf = []
for segment_set in utils.chunks(buf, 50):
self.client.put_trace_segments(
TraceSegmentDocuments=[
s.serialize() for s in segment_set])
def process(self, resources):
# Legacy
msg = self.data.get('msg')
msg = self.data.get('value') or msg
tag = self.data.get('tag', DEFAULT_TAG)
tag = self.data.get('key') or tag
# Support setting multiple tags in a single go with a mapping
tags = self.data.get('tags')
if tags is None:
tags = []
else:
tags = [{'Key': k, 'Value': v} for k, v in tags.items()]
if msg:
tags.append({'Key': tag, 'Value': msg})
batch_size = self.data.get('batch_size', self.batch_size)
with self.executor_factory(max_workers=self.concurrency) as w:
futures = []
for resource_set in utils.chunks(resources, size=batch_size):
futures.append(
w.submit(self.process_resource_set, resource_set, tags))
for f in as_completed(futures):
if f.exception():
self.log.error(
"Exception removing tags: %s on resources:%s \n %s" % (
tags, self.id_key, f.exception()))
def _process_bucket(self, b, p, key_log, w):
content_key = self.get_bucket_op(b, 'contents_key')
count = 0
for key_set in p:
count += len(key_set.get(content_key, []))
# Empty bucket check
if not content_key in key_set and not key_set['IsTruncated']:
return {'Bucket': b['Name'],
'Remediated': key_log.count,
'Count': count}
futures = []
for batch in chunks(key_set.get(content_key, []), size=100):
if not batch:
continue
futures.append(w.submit(self.process_chunk, batch, b))
for f in as_completed(futures):
if f.exception():
log.exception("Exception Processing bucket:%s key batch %s" % (
b['Name'], f.exception()))
continue
r = f.result()
if r:
key_log.add(r)
# Log completion at info level, progress at debug level
if key_set['IsTruncated']:
log.debug('Scan progress bucket:%s keys:%d remediated:%d ...',
b['Name'], count, key_log.count)
else:
log.info('Scan Complete bucket:%s keys:%d remediated:%d',
b['Name'], count, key_log.count)
return {'Bucket': b['Name'], 'Remediated': key_log.count, 'Count': count}
def add_tags_to_results(self, client, elbs):
"""
Gets the tags for the ELBs and adds them to
the result set.
"""
elb_names = [elb['LoadBalancerName'] for elb in elbs]
names_to_tags = {}
fn = partial(self.process_tags, client=client)
futures = []
with self.executor_factory(max_workers=3) as w:
# max 20 ELBs per call (API limitation)
for elb_names_chunk in chunks(elb_names, size=20):
futures.append(
w.submit(fn, elb_names_chunk))
for f in as_completed(futures):
if f.exception():
self.log.exception("Exception Processing ELB: %s" % (
f.exception()))
continue
r = f.result()
if r:
names_to_tags.update(r)
for elb in elbs:
elb['Tags'] = names_to_tags[elb['LoadBalancerName']]
def process(self, resources, event=None):
for batch in utils.chunks(resources, 500):
message = {'resources': batch,
'event': event,
'action': self.data,
'policy': self.manager.data}
self.send_data_message(message)
def process(self, resources, event=None):
if not resources:
return resources
client = local_session(self.manager.session_factory).client(
'health', region_name='us-east-1')
f = self.get_filter_parameters()
resource_map = {r[self.manager.get_model().id]: r for r in resources}
found = set()
seen = set()
for resource_set in chunks(resource_map.keys(), 100):
f['entityValues'] = resource_set
events = client.describe_events(filter=f)['events']
events = [e for e in events if e['arn'] not in seen]
entities = self.process_event(events)
event_map = {e['arn']: e for e in events}
for e in entities:
rid = e['entityValue']
if rid not in resource_map:
continue
resource_map[rid].setdefault(
'c7n:HealthEvent', []).append(event_map[e['eventArn']])
found.add(rid)
seen.update(event_map.keys())
return [resource_map[resource_id] for resource_id in found]
def augment(self, resources):
client = local_session(self.session_factory).client('health')
for resource_set in chunks(resources, 10):
event_map = {r['arn']: r for r in resource_set}
event_details = client.describe_event_details(
eventArns=event_map.keys())['successfulSet']
for d in event_details:
event_map[d['event']['arn']][
'Description'] = d['eventDescription']['latestDescription']
event_arns = [r['arn'] for r in resource_set
if r['eventTypeCategory'] != 'accountNotification']
if not event_arns:
continue
paginator = client.get_paginator('describe_affected_entities')
entities = list(itertools.chain(
*[p['entities']for p in paginator.paginate(
filter={'eventArns': event_arns})]))
for e in entities:
event_map[e.pop('eventArn')].setdefault(
'AffectedEntities', []).append(e)
return resources
def process(self, instances):
instances = self._filter_ec2_with_volumes(
self.filter_instance_state(instances))
if not len(instances):
return
client = utils.local_session(self.manager.session_factory).client('ec2')
failures = {}
# Play nice around aws having insufficient capacity...
for itype, t_instances in utils.group_by(
instances, 'InstanceType').items():
for izone, z_instances in utils.group_by(
t_instances, 'Placement.AvailabilityZone').items():
for batch in utils.chunks(z_instances, self.batch_size):
fails = self.process_instance_set(client, batch, itype, izone)
if fails:
failures["%s %s" % (itype, izone)] = [i['InstanceId'] for i in batch]
if failures:
fail_count = sum(map(len, failures.values()))
msg = "Could not start %d of %d instances %s" % (
fail_count, len(instances),
utils.dumps(failures))
self.log.warning(msg)
raise RuntimeError(msg)
def _datapipeline_info(pipes, session_factory, executor_factory, retry):
def process_tags(pipe_set):
client = local_session(session_factory).client('datapipeline')
pipe_map = {pipe['id']: pipe for pipe in pipe_set}
while True:
try:
results = retry(
client.describe_pipelines,
pipelineIds=list(pipe_map.keys()))
break
except ClientError as e:
if e.response['Error']['Code'] != 'PipelineNotFound':
raise
msg = e.response['Error']['Message']
_, lb_name = msg.strip().rsplit(' ', 1)
pipe_map.pop(lb_name)
if not pipe_map:
results = {'TagDescriptions': []}
break
continue
for pipe_desc in results['pipelineDescriptionList']:
pipe = pipe_map[pipe_desc['pipelineId']]
pipe['tags'] = pipe_desc['tags']
for field in pipe_desc['fields']:
key = field['key']
if not key.startswith('@'):
continue
pipe[key[1:]] = field['stringValue']
with executor_factory(max_workers=2) as w:
return list(w.map(process_tags, chunks(pipes, 20)))
def process(self, resources):
if self.data['target_region'] == self.manager.config.region:
self.log.warning(
"Source and destination region are the same, skipping copy")
return
for resource_set in chunks(resources, 20):
self.process_resource_set(resource_set)
def execute_in_parallel(resources, event, execution_method, executor_factory, log,
max_workers=constants.DEFAULT_MAX_THREAD_WORKERS,
chunk_size=constants.DEFAULT_CHUNK_SIZE):
futures = []
results = []
exceptions = []
if ThreadHelper.disable_multi_threading:
try:
result = execution_method(resources, event)
if result:
results.extend(result)
except Exception as e:
exceptions.append(e)
else:
with executor_factory(max_workers=max_workers) as w:
for resource_set in chunks(resources, chunk_size):
futures.append(w.submit(execution_method, resource_set, event))
for f in as_completed(futures):
if f.exception():
log.error(
"Execution failed with error: %s" % f.exception())
exceptions.append(f.exception())
else:
result = f.result()
if result:
results.extend(result)
return results, list(set(exceptions))
def load_manifest_file(client, bucket, schema, versioned, ifilters, key_info):
"""Given an inventory csv file, return an iterator over keys
"""
# To avoid thundering herd downloads, we do an immediate yield for
# interspersed i/o
yield None
# Inline these values to avoid the local var lookup, they are constants
# rKey = schema['Key'] # 1
# rIsLatest = schema['IsLatest'] # 3
# rVersionId = schema['VersionId'] # 2
with tempfile.NamedTemporaryFile() as fh:
client.download_fileobj(Bucket=bucket, Key=key_info['key'], Fileobj=fh)
fh.seek(0)
reader = csv.reader(gzip.GzipFile(fileobj=fh, mode='r'))
for key_set in chunks(reader, 1000):
keys = []
for kr in key_set:
k = kr[1]
if inventory_filter(ifilters, schema, kr):
continue
k = unquote_plus(k)
if versioned:
if kr[3] == 'true':
keys.append((k, kr[2], True))
else:
keys.append((k, kr[2]))
else:
keys.append(k)
yield keys
def process_resource_set(self, resource_set):
client = self.manager.session_factory(
region=self.data['target_region']).client('ec2')
if self.data['target_region'] != self.manager.config.region:
cross_region = True
params = {}
params['Encrypted'] = self.data.get('encrypted', True)
if params['Encrypted']:
params['KmsKeyId'] = self.data['target_key']
for snapshot_set in chunks(resource_set, 5):
for r in snapshot_set:
snapshot_id = client.copy_snapshot(
SourceRegion=self.manager.config.region,
SourceSnapshotId=r['SnapshotId'],
Description=r.get('Description', ''),
**params)['SnapshotId']
client.create_tags(
Resources=[snapshot_id],
Tags=r['Tags'])
r['CopiedSnapshot'] = snapshot_id
if not cross_region or len(snapshot_set) < 5:
continue
copy_ids = [r['CopiedSnapshot'] for r in snapshot_set]
self.log.debug(
"Waiting on cross-region snapshot copy %s", ",".join(copy_ids))
waiter = client.get_waiter('snapshot_completed')
waiter.wait(SnapshotIds=copy_ids)
self.log.debug(
"Cross region copy complete %s", ",".join(copy_ids))
def get_elb_bucket_locations(self):
session = local_session(self.manager.session_factory)
client = session.client('elb')
# Try to use the cache if it exists
elbs = self.manager._cache.get(
{'region': self.manager.config.region, 'resource': 'elb'})
# Sigh, post query refactor reuse, we can't save our cache here
# as that resource manager does extra lookups on tags. Not
# worth paginating, since with cache usage we have full set in
# mem.
if elbs is None:
p = client.get_paginator('describe_load_balancers')
results = p.paginate()
elbs = results.build_full_result().get(
'LoadBalancerDescriptions', ())
self.log.info("Queried %d elbs", len(elbs))
else:
self.log.info("Using %d cached elbs", len(elbs))
get_elb_attrs = functools.partial(
_query_elb_attrs, self.manager.session_factory)
with self.executor_factory(max_workers=2) as w:
futures = []
for elb_set in chunks(elbs, 100):
futures.append(w.submit(get_elb_attrs, elb_set))
for f in as_completed(futures):
if f.exception():
log.error("Error while scanning elb log targets: %s" % (
f.exception()))
continue
for tgt in f.result():
yield tgt
def process(self, resources):
# Move this to policy? / no resources bypasses actions?
if not len(resources):
return
msg_tmpl = self.data.get(
'msg',
'Resource does not meet policy: {op}@{action_date}')
op = self.data.get('op', 'stop')
tag = self.data.get('tag', DEFAULT_TAG)
date = self.data.get('days', 4)
n = datetime.now(tz=tzutc())
action_date = n + timedelta(days=date)
msg = msg_tmpl.format(
op=op, action_date=action_date.strftime('%Y/%m/%d'))
self.log.info("Tagging %d resources for %s on %s" % (
len(resources), op, action_date.strftime('%Y/%m/%d')))
tags = [{'Key': tag, 'Value': msg}]
with self.executor_factory(max_workers=2) as w:
futures = []
for resource_set in utils.chunks(resources, size=200):
futures.append(
w.submit(self.process_resource_set, resource_set, tags))
for f in as_completed(futures):
if f.exception():
self.log.error(
"Exception tagging resource set: %s \n %s" % (
tags, f.exception()))
def process(self, resources):
stopped_instances = self.filter_instance_state(
resources, ('stopped',))
running_instances = self.filter_instance_state(
resources, ('running',))
if self.data.get('restart') and running_instances:
Stop({'terminate-ephemeral': False},
self.manager).process(running_instances)
client = utils.local_session(
self.manager.session_factory).client('ec2')
waiter = client.get_waiter('instance_stopped')
try:
waiter.wait(
InstanceIds=[r['InstanceId'] for r in running_instances])
except ClientError as e:
self.log.exception(
"Exception stopping instances for resize:\n %s" % e)
for instance_set in utils.chunks(itertools.chain(
stopped_instances, running_instances), 20):
self.process_resource_set(instance_set)
if self.data.get('restart') and running_instances:
client.start_instances(
InstanceIds=[i['InstanceId'] for i in running_instances])
return list(itertools.chain(stopped_instances, running_instances))
def resources(self, query=None):
client = local_session(self.manager.session_factory).client('config')
paginator = client.get_paginator('list_discovered_resources')
pages = paginator.paginate(
resourceType=self.manager.get_model().config_type)
results = []
with self.manager.executor_factory(max_workers=5) as w:
ridents = pager(pages, self.retry)
resource_ids = [
r['resourceId'] for r in ridents.get('resourceIdentifiers', ())]
self.manager.log.debug(
"querying %d %s resources",
len(resource_ids),
self.manager.__class__.__name__.lower())
for resource_set in chunks(resource_ids, 50):
futures = []
futures.append(w.submit(self.get_resources, resource_set))
for f in as_completed(futures):
if f.exception():
self.manager.log.error(
"Exception getting resources from config \n %s" % (
f.exception()))
results.extend(f.result())
return results
def _elb_tags(elbs, session_factory, executor_factory, retry):
def process_tags(elb_set):
client = local_session(session_factory).client('elb')
elb_map = {elb['LoadBalancerName']: elb for elb in elb_set}
while True:
try:
results = retry(
client.describe_tags,
LoadBalancerNames=elb_map.keys())
break
except ClientError as e:
if e.response['Error']['Code'] != 'LoadBalancerNotFound':
raise
msg = e.response['Error']['Message']
_, lb_name = msg.strip().rsplit(' ', 1)
elb_map.pop(lb_name)
if not elb_map:
results = {'TagDescriptions': []}
break
continue
for tag_desc in results['TagDescriptions']:
elb_map[tag_desc['LoadBalancerName']]['Tags'] = tag_desc['Tags']
with executor_factory(max_workers=2) as w:
list(w.map(process_tags, chunks(elbs, 20)))
def process_account(account_info):
"""Scan all buckets in an account and schedule processing"""
log = logging.getLogger('salactus.bucket-iterator')
log.info("processing account %s", account_info)
session = get_session(account_info)
client = session.client('s3', config=s3config)
buckets = client.list_buckets()['Buckets']
connection.hset(
'bucket-accounts', account_info['name'], json.dumps(account_info))
for b in buckets:
connection.hset(
'bucket-ages', bucket_id(account_info, b['Name']),
b['CreationDate'].isoformat())
account_buckets = account_info.pop('buckets', None)
buckets = [n['Name'] for n in buckets
if not account_buckets or
n['Name'] in account_buckets]
account_not_buckets = account_info.pop('not-buckets', None)
buckets = [n for n in buckets
if not account_not_buckets or
n not in account_not_buckets]
log.info("processing %d buckets in account %s",
len(buckets), account_info['name'])
for bucket_set in chunks(buckets, 50):
invoke(process_bucket_set, account_info, bucket_set)
def process(self, resources):
if self.data['target_region'] == self.manager.config.region:
self.log.info(
"Source and destination region are the same, skipping")
return
with self.executor_factory(max_workers=2) as w:
list(w.map(self.process_resource_set, chunks(resources, 20)))
def _process_bucket(self, b, p, key_log, w):
count = 0
for key_set in p:
keys = self.get_keys(b, key_set)
count += len(keys)
futures = []
for batch in chunks(keys, size=100):
if not batch:
continue
futures.append(w.submit(self.process_chunk, batch, b))
for f in as_completed(futures):
if f.exception():
log.exception("Exception Processing bucket:%s key batch %s" % (
b['Name'], f.exception()))
continue
r = f.result()
if r:
key_log.add(r)
# Log completion at info level, progress at debug level
if key_set['IsTruncated']:
log.debug('Scan progress bucket:%s keys:%d remediated:%d ...',
b['Name'], count, key_log.count)
else:
log.info('Scan Complete bucket:%s keys:%d remediated:%d',
b['Name'], count, key_log.count)
b['KeyScanCount'] = count
b['KeyRemediated'] = key_log.count
return {
'Bucket': b['Name'], 'Remediated': key_log.count, 'Count': count}
def bulk_invoke(func, args, nargs):
"""Bulk invoke a function via queues
Uses internal implementation details of rq.
"""
# for comparison, simplest thing that works
# for i in nargs:
# argv = list(args)
# argv.append(i)
# func.delay(*argv)
# some variances between cpy and pypy, sniff detect
for closure in func.delay.func_closure:
if getattr(closure.cell_contents, 'queue', None):
ctx = closure.cell_contents
break
q = Queue(ctx.queue, connection=connection)
argv = list(args)
argv.append(None)
job = Job.create(
func, args=argv, connection=connection,
description="bucket-%s" % func.func_name,
origin=q.name, status=JobStatus.QUEUED, timeout=ctx.timeout,
result_ttl=0, ttl=ctx.ttl)
for n in chunks(nargs, 100):
job.created_at = datetime.utcnow()
with connection.pipeline() as pipe:
for s in n:
argv[-1] = s
job._id = unicode(uuid4())
job.args = argv
q.enqueue_job(job, pipeline=pipe)
pipe.execute()
def process(self, resources, event=None):
client = utils.local_session(
self.manager.session_factory).client('lambda')
params = dict(FunctionName=self.data['function'])
if self.data.get('qualifier'):
params['Qualifier'] = self.data['Qualifier']
if self.data.get('async', True):
params['InvocationType'] = 'Event'
payload = {
'version': VERSION,
'event': event,
'action': self.data,
'policy': self.manager.data}
results = []
for resource_set in utils.chunks(resources, self.data.get('batch_size', 250)):
payload['resources'] = resource_set
params['Payload'] = utils.dumps(payload)
result = client.invoke(**params)
result['Payload'] = result['Payload'].read()
results.append(result)
return results
def process(self, resources, event=None):
days = self.data.get('days', 14)
duration = timedelta(days)
self.metric = self.data['name']
self.end = datetime.utcnow()
self.start = self.end - duration
self.period = int(self.data.get('period', duration.total_seconds()))
self.statistics = self.data.get('statistics', 'Average')
self.model = self.manager.get_model()
self.op = OPERATORS[self.data.get('op', 'less-than')]
self.value = self.data['value']
ns = self.data.get('namespace')
if not ns:
ns = getattr(self.model, 'default_namespace', None)
if not ns:
ns = self.DEFAULT_NAMESPACE[self.model.service]
self.namespace = ns
self.log.debug("Querying metrics for %d", len(resources))
matched = []
with self.executor_factory(max_workers=3) as w:
futures = []
for resource_set in chunks(resources, 50):
futures.append(
w.submit(self.process_resource_set, resource_set))
for f in as_completed(futures):
if f.exception():
self.log.warning(
"CW Retrieval error: %s" % f.exception())
continue
matched.extend(f.result())
return matched
def initialize(self, asgs):
"""Get launch configs for the set of asgs"""
config_names = set()
for a in asgs:
config_names.add(a['LaunchConfigurationName'])
session = local_session(self.manager.session_factory)
client = session.client('autoscaling')
self.configs = {}
if len(asgs) > 50 and self.manager._cache.load():
configs = self.manager._cache.get(
{'region': self.manager.config.region,
'resource': 'launch-config'})
if configs:
self.log.info("Using cached configs")
self.configs = {
cfg['LaunchConfigurationName']: cfg for cfg in configs}
return
self.log.debug("querying %d launch configs" % len(config_names))
for cfg_set in chunks(config_names, 50):
for cfg in client.describe_launch_configurations(
LaunchConfigurationNames=cfg_set)['LaunchConfigurations']:
self.configs[cfg['LaunchConfigurationName']] = cfg
if len(asgs) > 50:
self.manager._cache.save(
{'region': self.manager.config.region,
'resource': 'launch-config'},
self.configs.values())
def process(self, resources):
client = local_session(
self.manager.session_factory).client('cloudwatch')
for resource_set in chunks(resources, size=100):
self.manager.retry(
client.delete_alarms,
AlarmNames=[r['AlarmName'] for r in resource_set])
def process_cluster_resources(self, client, cluster_id, services):
results = []
for service_set in chunks(services, self.manager.chunk_size):
results.extend(
client.describe_services(
cluster=cluster_id,
services=service_set).get('services', []))
return results
def process(self, resources):
if self.attr_filter:
resources = self.filter_resources(resources)
m = self.manager.get_model()
session = local_session(self.manager.session_factory)
client = self.get_client(session, m)
for resource_set in chunks(resources, self.chunk_size):
self.process_resource_set(client, m, resource_set)
def create_elb_active_attributes_tuples(self, elb_policy_tuples):
"""
creates a list of tuples for all attributes that are marked
as "true" in the load balancer's polices, e.g.
(myelb,['Protocol-SSLv1','Protocol-SSLv2'])
"""
active_policy_attribute_tuples = []
with self.executor_factory(max_workers=3) as w:
futures = []
for elb_policy_set in chunks(elb_policy_tuples, 50):
futures.append(
w.submit(self.process_elb_policy_set, elb_policy_set))
for f in as_completed(futures):
if f.exception():
self.log.error(
"Exception processing elb policies \n %s" %
(f.exception()))
continue
for elb_policies in f.result():
active_policy_attribute_tuples.append(elb_policies)
return active_policy_attribute_tuples
def bulk_invoke(func, args, nargs):
"""Bulk invoke a function via queues
Uses internal implementation details of rq.
"""
ctx = func.delay.func_closure[-1].cell_contents
q = Queue(ctx.queue, connection=connection)
argv = list(args)
argv.append(None)
job = Job.create(
func, args=argv, connection=connection,
description="bucket-%s" % func.func_name,
origin=q.name, status=JobStatus.QUEUED, timeout=None,
result_ttl=500, ttl=ctx.ttl)
for n in chunks(nargs, 100):
job.created_at = datetime.utcnow()
with connection.pipeline() as pipe:
for s in n:
argv[-1] = s
job._id = unicode(uuid4())
job.args = argv
q.enqueue_job(job, pipeline=pipe)
def process(self, snapshots):
self.image_snapshots = snaps = set()
# Be careful re image snapshots, we do this by default
# to keep things safe by default, albeit we'd get an error
# if we did try to delete something associated to an image.
if self.data.get('skip-ami-snapshots', True):
# Auto filter ami referenced snapshots, build map
c = local_session(self.manager.session_factory).client('ec2')
for i in c.describe_images(Owners=['self'])['Images']:
for dev in i.get('BlockDeviceMappings'):
if 'Ebs' in dev:
snaps.add(dev['Ebs']['SnapshotId'])
log.info("Deleting %d snapshots", len(snapshots))
with self.executor_factory(max_workers=3) as w:
futures = []
for snapshot_set in chunks(reversed(snapshots), size=50):
futures.append(
w.submit(self.process_snapshot_set, snapshot_set))
for f in as_completed(futures):
if f.exception():
self.log.error(
"Exception deleting snapshot set \n %s" % (
f.exception()))
def process(self, snapshots):
self.image_snapshots = set()
# Be careful re image snapshots, we do this by default
# to keep things safe by default, albeit we'd get an error
# if we did try to delete something associated to an image.
pre = len(snapshots)
snapshots = list(filter(None, _filter_ami_snapshots(self, snapshots)))
post = len(snapshots)
log.info("Deleting %d snapshots, auto-filtered %d ami-snapshots",
post, pre - post)
client = local_session(self.manager.session_factory).client('ec2')
with self.executor_factory(max_workers=2) as w:
futures = []
for snapshot_set in chunks(reversed(snapshots), size=50):
futures.append(
w.submit(self.process_snapshot_set, client, snapshot_set))
for f in as_completed(futures):
if f.exception():
self.log.error(
"Exception deleting snapshot set \n %s" % (
f.exception()))
return snapshots
def _describe_route53_tags(model, resources, session_factory, executor_factory,
retry):
def process_tags(resources):
client = local_session(session_factory).client('route53')
resource_map = {}
for r in resources:
k = r[model.id]
if "hostedzone" in k:
k = k.split("/")[-1]
resource_map[k] = r
for resource_batch in chunks(list(resource_map.keys()), 10):
results = retry(client.list_tags_for_resources,
ResourceType=model.arn_type,
ResourceIds=resource_batch)
for resource_tag_set in results['ResourceTagSets']:
if ('ResourceId' in resource_tag_set
and resource_tag_set['ResourceId'] in resource_map):
resource_map[resource_tag_set['ResourceId']][
'Tags'] = resource_tag_set['Tags']
with executor_factory(max_workers=2) as w:
return list(w.map(process_tags, chunks(resources, 20)))
def get_elb_bucket_locations(self):
session = local_session(self.manager.session_factory)
client = session.client('elb')
# Try to use the cache if it exists
elbs = self.manager._cache.get({
'region': self.manager.config.region,
'resource': 'elb'
})
# Sigh, post query refactor reuse, we can't save our cache here
# as that resource manager does extra lookups on tags. Not
# worth paginating, since with cache usage we have full set in
# mem.
if elbs is None:
p = client.get_paginator('describe_load_balancers')
results = p.paginate()
elbs = results.build_full_result().get('LoadBalancerDescriptions',
())
self.log.info("Queried %d elbs", len(elbs))
else:
self.log.info("Using %d cached elbs", len(elbs))
get_elb_attrs = functools.partial(_query_elb_attrs,
self.manager.session_factory)
with self.executor_factory(max_workers=2) as w:
futures = []
for elb_set in chunks(elbs, 100):
futures.append(w.submit(get_elb_attrs, elb_set))
for f in as_completed(futures):
if f.exception():
log.error("Error while scanning elb log targets: %s" %
(f.exception()))
continue
for tgt in f.result():
yield tgt
def execute_in_parallel(resources,
event,
execution_method,
executor_factory,
log,
max_workers=constants.DEFAULT_MAX_THREAD_WORKERS,
chunk_size=constants.DEFAULT_CHUNK_SIZE,
**kwargs):
futures = []
results = []
exceptions = []
if ThreadHelper.disable_multi_threading:
try:
result = execution_method(resources, event, **kwargs)
if result:
results.extend(result)
except Exception as e:
exceptions.append(e)
else:
with executor_factory(max_workers=max_workers) as w:
for resource_set in chunks(resources, chunk_size):
futures.append(
w.submit(execution_method, resource_set, event,
**kwargs))
for f in as_completed(futures):
if f.exception():
log.error("Execution failed with error: %s" %
f.exception())
exceptions.append(f.exception())
else:
result = f.result()
if result:
results.extend(result)
return results, list(set(exceptions))
def _datapipeline_info(pipes, session_factory, executor_factory, retry):
client = local_session(session_factory).client('datapipeline')
def process_tags(pipe_set):
pipe_map = {pipe['id']: pipe for pipe in pipe_set}
while True:
try:
results = retry(client.describe_pipelines,
pipelineIds=list(pipe_map.keys()))
break
except ClientError as e:
if e.response['Error']['Code'] != 'PipelineNotFound':
raise
msg = e.response['Error']['Message']
_, lb_name = msg.strip().rsplit(' ', 1)
pipe_map.pop(lb_name)
if not pipe_map:
results = {'TagDescriptions': []}
break
continue
for pipe_desc in results['pipelineDescriptionList']:
pipe = pipe_map[pipe_desc['pipelineId']]
pipe['Tags'] = [{
'Key': t['key'],
'Value': t['value']
} for t in pipe_desc['tags']]
for field in pipe_desc['fields']:
key = field['key']
if not key.startswith('@'):
continue
pipe[key[1:]] = field['stringValue']
with executor_factory(max_workers=2) as w:
return list(w.map(process_tags, chunks(pipes, 20)))
def process(self, param_groups):
client = local_session(self.manager.session_factory).client('rds')
params = []
for param in self.data.get('params', []):
params.append({
'ParameterName':
param['name'],
'ParameterValue':
param['value'],
'ApplyMethod':
param.get('apply-method', 'immediate'),
})
for param_group in param_groups:
name = self.get_pg_name(param_group)
# Fetch the existing parameters for this DB, so we only try to change the ones that are
# different.
cur_params = self.get_current_params(client, name)
changed_params = []
for param in params:
param_name = param['ParameterName']
if (param_name not in cur_params
or cur_params[param_name]['ParameterValue'] !=
param['ParameterValue']):
changed_params.append(param)
# Can only do 20 elements at a time per docs, so if we have more than that we will
# break it into multiple requests: https://goo.gl/Z6oGNv
for param_set in chunks(changed_params, 5):
self.do_modify(client, name, param_set)
self.log.info(
'Modified RDS parameter group %s (%i parameters changed, %i unchanged)',
name, len(changed_params),
len(params) - len(changed_params))
def process_resource_set(self, resource_set):
client = self.manager.session_factory(
region=self.data['target_region']).client('ec2')
if self.data['target_region'] != self.manager.config.region:
cross_region = True
params = {}
params['Encrypted'] = self.data.get('encrypted', True)
if params['Encrypted']:
params['KmsKeyId'] = self.data['target_key']
for snapshot_set in chunks(resource_set, 5):
for r in snapshot_set:
snapshot_id = client.copy_snapshot(
SourceRegion=self.manager.config.region,
SourceSnapshotId=r['SnapshotId'],
Description=r.get('Description', ''),
**params)['SnapshotId']
client.create_tags(
Resources=[snapshot_id],
Tags=r['Tags'])
r['CopiedSnapshot'] = snapshot_id
if not cross_region or len(snapshot_set) < 5:
continue
copy_ids = [r['CopiedSnapshot'] for r in snapshot_set]
self.log.debug(
"Waiting on cross-region snapshot copy %s", ",".join(copy_ids))
waiter = client.get_waiter('snapshot_completed')
waiter.config.delay = 60
waiter.config.max_attempts = 60
waiter.wait(SnapshotIds=copy_ids)
self.log.debug(
"Cross region copy complete %s", ",".join(copy_ids))
def process(self, instances):
instances = self._filter_ec2_with_volumes(
self.filter_instance_state(instances))
if not len(instances):
return
client = utils.local_session(
self.manager.session_factory).client('ec2')
# Play nice around aws having insufficient capacity...
for itype, t_instances in utils.group_by(instances,
'InstanceType').items():
for izone, z_instances in utils.group_by(
t_instances, 'AvailabilityZone').items():
for batch in utils.chunks(z_instances, self.batch_size):
self.process_instance_set(client, batch, itype, izone)
# Raise an exception after all batches process
if self.exception:
if self.exception.response['Error']['Code'] not in (
'InsufficientInstanceCapacity'):
self.log.exception("Error while starting instances error %s",
self.exception)
raise self.exception
def process_account(account_info):
"""Scan all buckets in an account and schedule processing"""
log = logging.getLogger('salactus.bucket-iterator')
log.info("processing account %s", account_info)
session = get_session(account_info)
client = session.client('s3', config=s3config)
buckets = client.list_buckets()['Buckets']
connection.hset('bucket-accounts', account_info['name'],
json.dumps(account_info))
for b in buckets:
connection.hset('bucket-ages', bucket_id(account_info, b['Name']),
b['CreationDate'].isoformat())
account_buckets = account_info.pop('buckets', None)
buckets = [
n['Name'] for n in buckets
if not account_buckets or n['Name'] in account_buckets
]
log.info("processing %d buckets in account %s", len(buckets),
account_info['name'])
for bucket_set in chunks(buckets, 50):
invoke(process_bucket_set, account_info, bucket_set)
def process(self, resources, event=None):
params = dict(FunctionName=self.data['function'])
if self.data.get('qualifier'):
params['Qualifier'] = self.data['Qualifier']
if self.data.get('async', True):
params['InvocationType'] = 'Event'
config = Config(read_timeout=self.data.get('timeout', 90),
region_name=self.data.get('region', None))
client = utils.local_session(self.manager.session_factory).client(
'lambda', config=config)
alias = utils.get_account_alias_from_sts(
utils.local_session(self.manager.session_factory))
payload = {
'version': VERSION,
'event': event,
'account_id': self.manager.config.account_id,
'account': alias,
'region': self.manager.config.region,
'action': self.data,
'policy': self.manager.data
}
results = []
for resource_set in utils.chunks(resources,
self.data.get('batch_size', 250)):
payload['resources'] = resource_set
params['Payload'] = utils.dumps(payload)
result = client.invoke(**params)
result['Payload'] = result['Payload'].read()
if isinstance(result['Payload'], bytes):
result['Payload'] = result['Payload'].decode('utf-8')
results.append(result)
return results
def initialize(self, asgs):
"""Get launch configs for the set of asgs"""
config_names = set()
for a in asgs:
config_names.add(a['LaunchConfigurationName'])
session = local_session(self.manager.session_factory)
client = session.client('autoscaling')
self.configs = {}
if len(asgs) > 50 and self.manager._cache.load():
configs = self.manager._cache.get({
'region': self.manager.config.region,
'resource': 'launch-config'
})
if configs:
self.log.info("Using cached configs")
self.configs = {
cfg['LaunchConfigurationName']: cfg
for cfg in configs
}
return
self.log.debug("querying %d launch configs" % len(config_names))
for cfg_set in chunks(config_names, 50):
for cfg in client.describe_launch_configurations(
LaunchConfigurationNames=cfg_set)['LaunchConfigurations']:
self.configs[cfg['LaunchConfigurationName']] = cfg
if len(asgs) > 50:
self.manager._cache.save(
{
'region': self.manager.config.region,
'resource': 'launch-config'
}, self.configs.values())
def process(self, resources):
self.id_key = self.manager.get_model().id
# Move this to policy? / no resources bypasses actions?
if not len(resources):
return
msg_tmpl = self.data.get(
'msg',
'Resource does not meet policy: {op}@{action_date}')
op = self.data.get('op', 'stop')
tag = self.data.get('tag', DEFAULT_TAG)
date = self.data.get('days', 4)
n = datetime.now(tz=tzutc())
action_date = n + timedelta(days=date)
msg = msg_tmpl.format(
op=op, action_date=action_date.strftime('%Y/%m/%d'))
self.log.info("Tagging %d resources for %s on %s" % (
len(resources), op, action_date.strftime('%Y/%m/%d')))
tags = [{'Key': tag, 'Value': msg}]
with self.executor_factory(max_workers=2) as w:
futures = []
for resource_set in utils.chunks(resources, size=self.batch_size):
futures.append(
w.submit(self.process_resource_set, resource_set, tags))
for f in as_completed(futures):
if f.exception():
self.log.error(
"Exception tagging resource set: %s \n %s" % (
tags, f.exception()))
def process(self, resources):
self.id_key = self.manager.get_model().id
# Legacy
msg = self.data.get('msg')
msg = self.data.get('value') or msg
tag = self.data.get('tag', DEFAULT_TAG)
tag = self.data.get('key') or tag
# Support setting multiple tags in a single go with a mapping
tags = self.data.get('tags')
if tags is None:
tags = []
else:
tags = [{'Key': k, 'Value': v} for k, v in tags.items()]
if msg:
tags.append({'Key': tag, 'Value': msg})
batch_size = self.data.get('batch_size', self.batch_size)
with self.executor_factory(max_workers=self.concurrency) as w:
futures = {}
for resource_set in utils.chunks(resources, size=batch_size):
futures[w.submit(self.process_resource_set, resource_set,
tags)] = resource_set
for f in as_completed(futures):
if f.exception():
self.log.error(
"Exception removing tags: %s on resources:%s \n %s" %
(tags, ", ".join(
[r[self.id_key]
for r in resource_set]), f.exception()))
def process(self, resources, event=None):
client = local_session(self.manager.session_factory).client('ssm')
item_template = self.get_item_template()
resources = list(
sorted(resources,
key=operator.itemgetter(self.manager.resource_type.id)))
items = self.get_items(client, item_template)
if items:
# - Use a copy of the template as we'll be passing in status changes on updates.
# - The return resources will be those that we couldn't fit into updates
# to existing resources.
resources = self.update_items(client, items, dict(item_template),
resources)
item_ids = [i['OpsItemId'] for i in items[:5]]
for resource_set in chunks(resources, 100):
resource_arns = json.dumps([{
'arn': arn
} for arn in sorted(self.manager.get_arns(resource_set))])
item_template['OperationalData']['/aws/resources'] = {
'Type': 'SearchableString',
'Value': resource_arns
}
if items:
item_template['RelatedOpsItems'] = [{
'OpsItemId': item_ids[:5]
}]
try:
oid = client.create_ops_item(**item_template).get('OpsItemId')
item_ids.insert(0, oid)
except client.exceptions.OpsItemAlreadyExistsException:
pass
for r in resource_set:
r['c7n:opsitem'] = oid
def universal_augment(self, resources):
# Resource Tagging API Support
# https://docs.aws.amazon.com/awsconsolehelpdocs/latest/gsg/supported-resources.html
# Bail on empty set
if not resources:
return resources
# For global resources, tags don't populate in the get_resources call
# unless the call is being made to us-east-1
region = getattr(self.resource_type, 'global_resource',
None) and 'us-east-1' or self.region
client = utils.local_session(self.session_factory).client(
'resourcegroupstaggingapi', region_name=region)
# Lazy for non circular :-(
from c7n.query import RetryPageIterator
paginator = client.get_paginator('get_resources')
paginator.PAGE_ITERATOR_CLS = RetryPageIterator
rfetch = [r for r in resources if 'Tags' not in r]
for arn_resource_set in utils.chunks(zip(self.get_arns(rfetch), rfetch),
100):
arn_resource_map = dict(arn_resource_set)
resource_tag_results = client.get_resources(
ResourceARNList=list(arn_resource_map.keys())).get(
'ResourceTagMappingList', ())
resource_tag_map = {
r['ResourceARN']: r['Tags']
for r in resource_tag_results
}
for arn, r in arn_resource_map.items():
r['Tags'] = resource_tag_map.get(arn, [])
return resources
def process(self, resources):
session = local_session(self.manager.session_factory)
m = self.manager.get_model()
client = session.client(m.group, m.version)
for resource_set in chunks(resources, self.chunk_size):
self.process_resource_set(client, resource_set)
def put_evaluations(self, client, token, evaluations):
for eval_set in utils.chunks(evaluations, 100):
self.policy.resource_manager.retry(client.put_evaluations,
Evaluations=eval_set,
ResultToken=token)
def process(self, resources):
for resource_set in chunks(resources, 50):
self.process_resource_set(resource_set)
def process(self, resources, event=None):
results = []
filtered = []
attached = []
stats = Counter()
marker_date = parse_date('2016-11-01T00:00:00+00:00')
# Filter volumes
for r in resources:
# unsupported type
if r['VolumeType'] == 'standard':
stats['vol-type'] += 1
filtered.append(r['VolumeId'])
continue
# unattached are easy
if not r.get('Attachments'):
results.append(r)
continue
# check for attachment date older then supported date
if r['Attachments'][0]['AttachTime'] < marker_date:
stats['attach-time'] += 1
filtered.append(r['VolumeId'])
continue
attached.append(r)
# Filter volumes attached to unsupported instance types
ec2 = self.manager.get_resource_manager('ec2')
instance_map = {}
for v in attached:
instance_map.setdefault(
v['Attachments'][0]['InstanceId'], []).append(v)
instances = ec2.get_resources(list(instance_map.keys()))
for i in instances:
if i['InstanceType'] in self.older_generation:
stats['instance-type'] += len(instance_map[i['InstanceId']])
filtered.extend([v['VolumeId'] for v in instance_map.pop(i['InstanceId'])])
else:
results.extend(instance_map.pop(i['InstanceId']))
# Filter volumes that are currently under modification
client = local_session(self.manager.session_factory).client('ec2')
modifying = set()
for vol_set in chunks(list(results), 200):
vol_ids = [v['VolumeId'] for v in vol_set]
mutating = client.describe_volumes_modifications(
Filters=[
{'Name': 'volume-id',
'Values': vol_ids},
{'Name': 'modification-state',
'Values': ['modifying', 'optimizing', 'failed']}])
for vm in mutating.get('VolumesModifications', ()):
stats['vol-mutation'] += 1
filtered.append(vm['VolumeId'])
modifying.add(vm['VolumeId'])
self.log.debug(
"filtered %d of %d volumes due to %s",
len(filtered), len(resources), sorted(stats.items()))
return [r for r in results if r['VolumeId'] not in modifying]
def enable_region(master_info, accounts_config, executor, message, region,
enable_email_notification):
master_session = get_session(master_info.get('role'),
'c7n-guardian',
master_info.get('profile'),
region=region)
master_client = master_session.client('guardduty')
detector_id = get_or_create_detector_id(master_client)
if not detector_id:
# Couldn't get a detector in this region; perhaps region not opted-in
log.info(
f"""{region}\tCouldn't get or create GuardDuty detector with role
{master_info.get('role')} - Perhaps the region isn't enabled?"""
)
return
# The list gd_members contains active GuardDuty Member Accounts
# If the account is in the specified account list (.yml) and not in gd_members,
# we'll ensure it's emabled
results = master_client.get_paginator('list_members').paginate(
DetectorId=detector_id, OnlyAssociated="True")
gd_members = results.build_full_result().get('Members', ())
gd_member_ids = {m['AccountId'] for m in gd_members}
# Build this the ugly way to ensure we break if there's an unexpected RelationshipStatus;
# rather break than silently pass
active_ids = []
invited_ids = []
suspended_ids = []
resigned_ids = []
removed_ids = []
for m in gd_members:
if m['RelationshipStatus'] == 'Enabled':
active_ids.append(m['AccountId'])
elif m['RelationshipStatus'] in ['Invited']:
invited_ids.append(m['AccountId'])
elif m['RelationshipStatus'] == 'Disabled':
suspended_ids.append(m['AccountId'])
elif m['RelationshipStatus'] == 'Removed':
# The GD detector has been entirely removed - re-invite & accept
removed_ids.append(m['AccountId'])
elif m['RelationshipStatus'] == 'Resigned':
# If member is Resigned = member has switched off the master (GD still enabled though)
# We should not re-invite, but should re-accept
resigned_ids.append(m['AccountId'])
else:
raise Exception(f'''GuardDuty member account {m["AccountId"]} had
unknown RelationshipStatus
"{m["RelationshipStatus"]}" - bailing''')
# Filter by accounts under consideration per config and cli flags
suspended_ids = {
a['account_id']
for a in accounts_config['accounts']
if a['account_id'] in suspended_ids
}
if suspended_ids:
unprocessed = master_client.start_monitoring_members(
DetectorId=detector_id,
AccountIds=list(suspended_ids)).get('UnprocessedAccounts')
if unprocessed:
log.warning(
f"""{region}\tUnprocessed accounts on re-start monitoring
{format_event(unprocessed)}""")
log.info(
"{region}\tRestarted monitoring on {len(suspended_ids)} accounts")
accounts_not_members = [{
'AccountId': account['account_id'],
'Email': account['email']
} for account in accounts_config['accounts']
if account['account_id'] not in gd_member_ids]
if not accounts_not_members:
if not suspended_ids and not invited_ids and not resigned_ids and not removed_ids:
log.info(f"{region}\tAll accounts already enabled")
return list(active_ids)
if (len(accounts_not_members) + len(gd_member_ids)) > 1000:
raise ValueError(f"""{region}\tGuard Duty only supports 1000 member
accounts per master account""")
log.info(
f"{region}\tEnrolling {len(accounts_not_members)} accounts in guard duty"
)
unprocessed = []
for account_set in chunks(accounts_not_members, 25):
new_members = master_client.create_members(DetectorId=detector_id,
AccountDetails=account_set)
unprocessed.extend(new_members.get('UnprocessedAccounts', []))
# If the account was already a member, ignore
unprocessed = list(
filter(lambda x: x['Result'].find('already a membe') == -1,
unprocessed))
if unprocessed:
log.warning(f"""{region}\tAccounts were unprocessed - member create
{format_event(unprocessed)}""")
log.info(f"{region}\tInviting {len(accounts_not_members)} member accounts")
unprocessed = []
for account_set in chunks([
m for m in accounts_not_members
if not m['AccountId'] in invited_ids + resigned_ids
], 25):
params = {
'AccountIds': [m['AccountId'] for m in account_set],
'DetectorId': detector_id,
'DisableEmailNotification': not enable_email_notification,
}
if message:
params['Message'] = message
unprocessed.extend(
master_client.invite_members(**params).get('UnprocessedAccounts',
[]))
if unprocessed:
log.warning(f"""{region}\tAccounts were unprocessed invite-members
{format_event(unprocessed)}""")
accounts_not_members = [{
'AccountId': account['account_id'],
'Email': account['email']
} for account in accounts_config['accounts']
if account['account_id'] not in active_ids]
log.info(
f"{region}\tAccepting {len(accounts_not_members)} invitations in members"
)
with executor(max_workers=WORKER_COUNT) as w:
futures = {}
for a in accounts_config['accounts']:
if a == master_info:
continue
if a['account_id'] in active_ids:
continue
futures[w.submit(enable_account, a, master_info['account_id'],
region)] = a
for f in as_completed(futures):
a = futures[f]
if f.exception():
log.error(
f"{region}\tError processing account:{a['name']} error:{f.exception()}"
)
continue
if f.result():
log.info(
f"{region}\tEnabled guard duty on account:{a['name']}")
return accounts_not_members
def flush(self):
buf = self.buf
self.buf = []
for segment_set in utils.chunks(buf, 50):
self.client.put_trace_segments(
TraceSegmentDocuments=[s.serialize() for s in segment_set])
def process(self, resources):
client = local_session(self.manager.session_factory).client('ssm')
for resource_set in chunks(resources, 50):
self.process_resource_set(client, resource_set)
def enable_region(master_info, accounts_config, executor, message, region):
master_session = get_session(
master_info.get('role'), 'c7n-guardian',
master_info.get('profile'),
region=region)
master_client = master_session.client('guardduty')
detector_id = get_or_create_detector_id(master_client)
results = master_client.get_paginator(
'list_members').paginate(DetectorId=detector_id, OnlyAssociated="FALSE")
extant_members = results.build_full_result().get('Members', ())
extant_ids = {m['AccountId'] for m in extant_members}
# Find active members
active_ids = {m['AccountId'] for m in extant_members
if m['RelationshipStatus'] == 'Enabled'}
# Find invited members
invited_ids = {m['AccountId'] for m in extant_members
if m['RelationshipStatus'] == 'Invited'}
# Find extant members not currently enabled
suspended_ids = {m['AccountId'] for m in extant_members
if m['RelationshipStatus'] == 'Disabled'}
# Filter by accounts under consideration per config and cli flags
suspended_ids = {a['account_id'] for a in accounts_config['accounts']
if a['account_id'] in suspended_ids}
if suspended_ids:
unprocessed = master_client.start_monitoring_members(
DetectorId=detector_id,
AccountIds=list(suspended_ids)).get('UnprocessedAccounts')
if unprocessed:
log.warning(
"Region: %s Unprocessed accounts on re-start monitoring %s",
region, format_event(unprocessed))
log.info("Region: %s Restarted monitoring on %d accounts",
region, len(suspended_ids))
members = [{'AccountId': account['account_id'], 'Email': account['email']}
for account in accounts_config['accounts']
if account['account_id'] not in extant_ids]
if not members:
if not suspended_ids and not invited_ids:
log.info("Region:%s All accounts already enabled", region)
return list(active_ids)
if (len(members) + len(extant_ids)) > 1000:
raise ValueError(
("Region:%s Guard Duty only supports "
"1000 member accounts per master account") % (region))
log.info(
"Region:%s Enrolling %d accounts in guard duty", region, len(members))
unprocessed = []
for account_set in chunks(members, 25):
unprocessed.extend(master_client.create_members(
DetectorId=detector_id,
AccountDetails=account_set).get('UnprocessedAccounts', []))
if unprocessed:
log.warning(
"Region:%s accounts where unprocessed - member create\n %s",
region, format_event(unprocessed))
log.info("Region:%s Inviting %d member accounts", region, len(members))
unprocessed = []
for account_set in chunks(
[m for m in members if not m['AccountId'] in invited_ids], 25):
params = {'AccountIds': [m['AccountId'] for m in account_set],
'DetectorId': detector_id}
if message:
params['Message'] = message
unprocessed.extend(master_client.invite_members(
**params).get('UnprocessedAccounts', []))
if unprocessed:
log.warning(
"Region:%s accounts where unprocessed invite-members\n %s",
region, format_event(unprocessed))
members = [{'AccountId': account['account_id'], 'Email': account['email']}
for account in accounts_config['accounts']
if account['account_id'] not in active_ids]
log.info("Region:%s Accepting %d invitations in members", region, len(members))
with executor(max_workers=WORKER_COUNT) as w:
futures = {}
for a in accounts_config['accounts']:
if a == master_info:
continue
if a['account_id'] in active_ids:
continue
futures[w.submit(enable_account, a, master_info['account_id'], region)] = a
for f in as_completed(futures):
a = futures[f]
if f.exception():
log.error("Region:%s Error processing account:%s error:%s",
region, a['name'], f.exception())
continue
if f.result():
log.info('Region:%s Enabled guard duty on account:%s',
region, a['name'])
return members
def process_keyset(bid, key_set):
account, bucket = bid.split(':', 1)
region = connection.hget('bucket-regions', bid)
versioned = bool(int(connection.hget('bucket-versions', bid)))
account_info = json.loads(connection.hget('bucket-accounts', account))
visitors = get_key_visitors(account_info)
object_reporting = account_info.get('object-reporting')
session = get_session(account_info)
patch_ssl()
s3 = session.client('s3', region_name=region, config=s3config)
error_count = sesserr = connerr = enderr = missing_count = 0
throttle_count = denied_count = remediation_count = 0
key_count = len(key_set)
start_time = time.time()
objects = {v.visitor_name: [] for v in visitors}
objects['objects_denied'] = []
with bucket_ops(bid, 'key'):
#MainThreadExecutor.async = False
#with MainThreadExecutor(max_workers=10) as w:
with ThreadPoolExecutor(max_workers=10) as w:
futures = {}
for kchunk in chunks(key_set, 100):
for v in visitors:
processor = (versioned and v.process_version
or v.process_key)
futures[w.submit(process_key_chunk, s3,
bucket, kchunk, processor,
bool(object_reporting))] = v.visitor_name
for f in as_completed(futures):
if f.exception():
log.warning("key error: %s", f.exception())
error_count += 1
continue
stats = f.result()
remediation_count += stats['remediated']
denied_count += stats['denied']
missing_count += stats['missing']
throttle_count += stats['throttle']
sesserr += stats['session']
enderr += stats['endpoint']
connerr += stats['connection']
if object_reporting:
vname = futures[f]
objects[vname].extend(stats['objects'])
objects['objects_denied'].extend(stats['objects_denied'])
with connection.pipeline() as p:
if remediation_count:
p.hincrby('keys-matched', bid, remediation_count)
if denied_count:
p.hincrby('keys-denied', bid, denied_count)
if missing_count:
p.hincrby('keys-missing', bid, missing_count)
if throttle_count:
p.hincrby('keys-throttled', bid, throttle_count)
if sesserr:
p.hincrby('keys-sesserr', bid, sesserr)
if connerr:
p.hincrby('keys-connerr', bid, connerr)
if enderr:
p.hincrby('keys-enderr', bid, enderr)
if error_count:
p.hincrby('keys-error', bid, error_count)
p.hincrby('keys-scanned', bid, key_count)
# track count again as we reset metrics period
p.hincrby('keys-count', bid, key_count)
p.hincrby('keys-time', bid, int(time.time() - start_time))
p.execute()
# write out object level info
if object_reporting:
publish_object_records(bid, objects, object_reporting)
# trigger some mem collection
if getattr(sys, 'pypy_version_info', None):
gc.collect()
def enable(config, master, tags, accounts, debug, message, region):
"""enable guard duty on a set of accounts"""
accounts_config, master_info, executor = guardian_init(
config, debug, master, accounts, tags)
master_session = get_session(master_info.get('role'),
'c7n-guardian',
master_info.get('profile'),
region=region)
master_client = master_session.client('guardduty')
detector_id = get_or_create_detector_id(master_client)
extant_members = master_client.list_members(DetectorId=detector_id).get(
'Members', ())
extant_ids = {m['AccountId'] for m in extant_members}
# Find extant members not currently enabled
suspended_ids = {
m['AccountId']
for m in extant_members if m['RelationshipStatus'] == 'Disabled'
}
# Filter by accounts under consideration per config and cli flags
suspended_ids = {
a['account_id']
for a in accounts_config['accounts']
if a['account_id'] in suspended_ids
}
if suspended_ids:
unprocessed = master_client.start_monitoring_members(
DetectorId=detector_id,
AccountIds=list(suspended_ids)).get('UnprocessedAccounts')
if unprocessed:
log.warning("Unprocessed accounts on re-start monitoring %s" %
(format_event(unprocessed)))
log.info("Restarted monitoring on %d accounts" % (len(suspended_ids)))
members = [{
'AccountId': account['account_id'],
'Email': account['email']
} for account in accounts_config['accounts']
if account['account_id'] not in extant_ids]
if not members:
if not suspended_ids:
log.info("All accounts already enabled")
return
if (len(members) + len(extant_ids)) > 1000:
raise ValueError(
"Guard Duty only supports 1000 member accounts per master account")
log.info("Enrolling %d accounts in guard duty" % len(members))
log.info("Creating member accounts:%d region:%s", len(members), region)
unprocessed = []
for account_set in chunks(members, 25):
unprocessed.extend(
master_client.create_members(DetectorId=detector_id,
AccountDetails=account_set).get(
'UnprocessedAccounts', []))
if unprocessed:
log.warning("Following accounts where unprocessed\n %s" %
format_event(unprocessed))
log.info("Inviting member accounts")
params = {
'AccountIds': [m['AccountId'] for m in members],
'DetectorId': detector_id
}
if message:
params['Message'] = message
unprocessed = master_client.invite_members(
**params).get('UnprocessedAccounts')
if unprocessed:
log.warning("Following accounts where unprocessed\n %s" %
format_event(unprocessed))
log.info("Accepting invitations")
with executor(max_workers=WORKER_COUNT) as w:
futures = {}
for a in accounts_config['accounts']:
if a == master_info:
continue
if a['account_id'] in extant_ids:
continue
futures[w.submit(enable_account, a, master_info['account_id'],
region)] = a
for f in as_completed(futures):
a = futures[f]
if f.exception():
log.error("Error processing account:%s error:%s", a['name'],
f.exception())
continue
if f.result():
log.info('Enabled guard duty on account:%s' % a['name'])
def process_bucket_partitions(
account_info, bucket, prefix_set=('',), partition='/',
strategy=None, limit=4):
"""Split up a bucket keyspace into smaller sets for parallel iteration.
"""
if strategy is None:
return detect_partition_strategy(account_info, bucket)
strategy = get_partition_strategy(account_info, bucket, strategy)
strategy.limit = limit
strategy.partition = partition
(contents_key,
contents_method,
continue_tokens) = BUCKET_OBJ_DESC[bucket['versioned']]
prefix_queue = strategy.initialize_prefixes(prefix_set)
keyset = []
bid = bucket_id(account_info, bucket['name'])
log.info("Process partition bid:%s strategy:%s delimiter:%s queue:%d limit:%d",
bid, strategy.__class__.__name__[0], partition, len(prefix_queue), limit)
session = get_session(account_info)
s3 = session.client('s3', region_name=bucket['region'], config=s3config)
def statm(prefix):
return "keyset:%d queue:%d prefix:%s bucket:%s size:%d" % (
len(keyset), len(prefix_queue), prefix, bid, bucket['keycount'])
while prefix_queue:
connection.hincrby('bucket-partition', bid, 1)
prefix = prefix_queue.pop()
if strategy.is_depth_exceeded(prefix):
log.info("Partition max depth reached, %s", statm(prefix))
invoke(process_bucket_iterator, account_info, bucket, prefix)
continue
method = getattr(s3, contents_method, None)
results = page_strip(method(
Bucket=bucket['name'], Prefix=prefix, Delimiter=partition),
bucket)
keyset.extend(results.get(contents_key, ()))
# As we probe we find keys, process any found
if len(keyset) > PARTITION_KEYSET_THRESHOLD:
log.info("Partition, processing keyset %s", statm(prefix))
invoke(
process_keyset, account_info, bucket,
page_strip({contents_key: keyset}, bucket))
keyset = []
strategy.find_partitions(prefix_queue, results)
# Do we have more than 1k keys at this level, continue iteration
continuation_params = {
k: results[k] for k in continue_tokens if k in results}
if continuation_params:
bp = int(connection.hget('bucket-partition', bid))
log.info("Partition has 1k keys, %s %s", statm(prefix), bp)
if not prefix_queue and bp < 5:
log.info("Recursive detection")
return detect_partition_strategy(account_info, bucket, prefix=prefix)
invoke(process_bucket_iterator,
account_info, bucket, prefix, delimiter=partition,
**continuation_params)
# If the queue get too deep, then go parallel
if len(prefix_queue) > PARTITION_QUEUE_THRESHOLD:
log.info("Partition add friends, %s", statm(prefix))
for s_prefix_set in chunks(
prefix_queue[PARTITION_QUEUE_THRESHOLD - 1:],
PARTITION_QUEUE_THRESHOLD - 1):
for s in list(s_prefix_set):
if strategy.is_depth_exceeded(prefix):
invoke(process_bucket_iterator,
account_info, bucket, s)
s_prefix_set.remove(s)
if not s_prefix_set:
continue
invoke(process_bucket_partitions,
account_info, bucket,
prefix_set=s_prefix_set, partition=partition,
strategy=strategy, limit=limit)
prefix_queue = prefix_queue[:PARTITION_QUEUE_THRESHOLD - 1]
if keyset:
invoke(process_keyset, account_info, bucket, {contents_key: keyset})
def test_chunks(self):
self.assertEqual(list(utils.chunks(range(100), size=50)),
[range(50), range(50, 100, 1)])
self.assertEqual(list(utils.chunks(range(1), size=50)), [range(1)])
self.assertEqual(list(utils.chunks(range(60), size=50)),
[range(50), range(50, 60, 1)])
