def send_email(to, subject, html_content, files=None,
dryrun=False, cc=None, bcc=None,
mime_subtype='mixed', **kwargs):
"""
Send an email with html content using sendgrid.
To use this plugin:
0. include sendgrid subpackage as part of your Airflow installation, e.g.,
pip install airflow[sendgrid]
1. update [email] backend in airflow.cfg, i.e.,
[email]
email_backend = airflow.contrib.utils.sendgrid.send_email
2. configure Sendgrid specific environment variables at all Airflow instances:
SENDGRID_MAIL_FROM={your-mail-from}
SENDGRID_API_KEY={your-sendgrid-api-key}.
"""
mail = Mail()
mail.from_email = Email(os.environ.get('SENDGRID_MAIL_FROM'))
mail.subject = subject
# Add the recipient list of to emails.
personalization = Personalization()
to = get_email_address_list(to)
for to_address in to:
personalization.add_to(Email(to_address))
if cc:
cc = get_email_address_list(cc)
for cc_address in cc:
personalization.add_cc(Email(cc_address))
if bcc:
bcc = get_email_address_list(bcc)
for bcc_address in bcc:
personalization.add_bcc(Email(bcc_address))
# Add custom_args to personalization if present
pers_custom_args = kwargs.get('personalization_custom_args', None)
if isinstance(pers_custom_args, dict):
for key in pers_custom_args.keys():
personalization.add_custom_arg(CustomArg(key, pers_custom_args[key]))
mail.add_personalization(personalization)
mail.add_content(Content('text/html', html_content))
categories = kwargs.get('categories', [])
for cat in categories:
mail.add_category(Category(cat))
# Add email attachment.
for fname in files or []:
basename = os.path.basename(fname)
attachment = Attachment()
with open(fname, "rb") as f:
attachment.content = base64.b64encode(f.read())
attachment.type = mimetypes.guess_type(basename)[0]
attachment.filename = basename
attachment.disposition = "attachment"
attachment.content_id = '<%s>' % basename
mail.add_attachment(attachment)
_post_sendgrid_mail(mail.get())
def execute(self, context):
s3_hook = S3Hook(self.aws_conn_id)
file_list = s3_hook.list_keys(
bucket_name=self.s3_bucket,
prefix=self.s3_key,
delimiter='/'
)
header = ''
body = ''
for file in file_list:
if 'header' in file:
header = s3_hook.read_key(
key=file,
bucket_name=self.s3_bucket
)
else:
body_text = s3_hook.read_key(
key=file,
bucket_name=self.s3_bucket
)
body = body + body_text
if header == '':
contents = body
else:
contents = header + body
SMTP_MAIL_FROM = configuration.conf.get('smtp', 'SMTP_MAIL_FROM')
to = get_email_address_list(self.to)
msg = MIMEMultipart(self.mime_subtype)
msg['Subject'] = self.subject
msg['From'] = SMTP_MAIL_FROM
msg['To'] = ", ".join(self.to)
recipients = self.to
if self.cc:
cc = get_email_address_list(self.cc)
msg['CC'] = ", ".join(cc)
recipients = recipients + cc
if self.bcc:
# don't add bcc in header
bcc = get_email_address_list(self.bcc)
recipients = recipients + bcc
msg['Date'] = formatdate(localtime=True)
mime_text = MIMEText(self.html_content, 'html', self.mime_charset)
msg.attach(mime_text)
filename = '{filename}{attachment_extension}'.format(filename=self.filename, attachment_extension=self.attachment_extension)
attachment = MIMEText(contents, 'plain')
attachment.add_header('Content-Disposition', 'attachment', filename=filename)
msg.attach(attachment)
send_MIME_email(SMTP_MAIL_FROM, recipients, msg)
def send_email(to, subject, html_content, files=None,
dryrun=False, cc=None, bcc=None,
mime_subtype='mixed', **kwargs):
"""
Send an email with html content using sendgrid.
To use this plugin:
0. include sendgrid subpackage as part of your Airflow installation, e.g.,
pip install airflow[sendgrid]
1. update [email] backend in airflow.cfg, i.e.,
[email]
email_backend = airflow.contrib.utils.sendgrid.send_email
2. configure Sendgrid specific environment variables at all Airflow instances:
SENDGRID_MAIL_FROM={your-mail-from}
SENDGRID_API_KEY={your-sendgrid-api-key}.
"""
mail = Mail()
mail.from_email = Email(os.environ.get('SENDGRID_MAIL_FROM'))
mail.subject = subject
# Add the recipient list of to emails.
personalization = Personalization()
to = get_email_address_list(to)
for to_address in to:
personalization.add_to(Email(to_address))
if cc:
cc = get_email_address_list(cc)
for cc_address in cc:
personalization.add_cc(Email(cc_address))
if bcc:
bcc = get_email_address_list(bcc)
for bcc_address in bcc:
personalization.add_bcc(Email(bcc_address))
mail.add_personalization(personalization)
mail.add_content(Content('text/html', html_content))
# Add custom_args to personalization if present
pers_custom_args = kwargs.get('personalization_custom_args', None)
if isinstance(pers_custom_args, dict):
for key in pers_custom_args.keys():
personalization.add_custom_arg(CustomArg(key, pers_custom_args[key]))
# Add email attachment.
for fname in files or []:
basename = os.path.basename(fname)
attachment = Attachment()
with open(fname, "rb") as f:
attachment.content = base64.b64encode(f.read())
attachment.type = mimetypes.guess_type(basename)[0]
attachment.filename = basename
attachment.disposition = "attachment"
attachment.content_id = '<%s>' % basename
mail.add_attachment(attachment)
_post_sendgrid_mail(mail.get())
def log_email_backend(to,
subject,
html_content,
files=None,
dryrun=False,
cc=None,
bcc=None,
mime_subtype='mixed',
mime_charset='utf-8',
**kwargs):
smtp_mail_from = configuration.conf.get('smtp', 'SMTP_MAIL_FROM')
to = get_email_address_list(to)
msg = MIMEMultipart(mime_subtype)
msg['Subject'] = subject
msg['From'] = smtp_mail_from
msg['To'] = ", ".join(to)
if cc:
cc = get_email_address_list(cc)
msg['CC'] = ", ".join(cc)
if bcc:
# don't add bcc in header
bcc = get_email_address_list(bcc)
msg['Date'] = formatdate(localtime=True)
mime_text = MIMEText(None, 'plain', mime_charset)
mime_text.replace_header('content-transfer-encoding', 'quoted-printable')
mime_text.set_payload(html_content)
msg.attach(mime_text)
for fname in files or []:
basename = os.path.basename(fname)
with open(fname, "rb") as f:
part = MIMEApplication(f.read(), Name=basename)
part[
'Content-Disposition'] = 'attachment; filename="%s"' % basename
part['Content-ID'] = '<%s>' % basename
msg.attach(part)
logger = logging.getLogger("plugin.log_email_backend")
logger.info("\n" + msg.as_string())
def log_email_backend(to, subject, html_content, files=None,
dryrun=False, cc=None, bcc=None,
mime_subtype='mixed', mime_charset='utf-8',
**kwargs):
smtp_mail_from = configuration.conf.get('smtp', 'SMTP_MAIL_FROM')
to = get_email_address_list(to)
msg = MIMEMultipart(mime_subtype)
msg['Subject'] = subject
msg['From'] = smtp_mail_from
msg['To'] = ", ".join(to)
if cc:
cc = get_email_address_list(cc)
msg['CC'] = ", ".join(cc)
if bcc:
# don't add bcc in header
bcc = get_email_address_list(bcc)
msg['Date'] = formatdate(localtime=True)
mime_text = MIMEText(None, 'plain', mime_charset)
mime_text.replace_header('content-transfer-encoding', 'quoted-printable')
mime_text.set_payload(html_content)
msg.attach(mime_text)
for fname in files or []:
basename = os.path.basename(fname)
with open(fname, "rb") as f:
part = MIMEApplication(
f.read(),
Name=basename
)
part['Content-Disposition'] = 'attachment; filename="%s"' % basename
part['Content-ID'] = '<%s>' % basename
msg.attach(part)
logger = logging.getLogger("plugin.log_email_backend")
logger.info("\n" + msg.as_string())
def test_get_email_address_colon_sep_string(self):
emails_string = '[email protected]; [email protected]'
self.assertEqual(
get_email_address_list(emails_string), EMAILS)
def test_get_email_address_invalid_type_in_iterable(self):
emails_list = ['*****@*****.**', 2]
with pytest.raises(TypeError):
get_email_address_list(emails_list)
def test_get_email_address_invalid_type(self):
emails_string = 1
with pytest.raises(TypeError):
get_email_address_list(emails_string)
def test_get_email_address_list(self):
emails_list = ['*****@*****.**', '*****@*****.**']
self.assertEqual(
get_email_address_list(emails_list), EMAILS)
def send_email(to, subject, html_content, files=None, dryrun=False, cc=None,
bcc=None, mime_subtype='mixed', sandbox_mode=False, **kwargs):
"""
Send an email with html content using sendgrid.
To use this plugin:
0. include sendgrid subpackage as part of your Airflow installation, e.g.,
pip install 'apache-airflow[sendgrid]'
1. update [email] backend in airflow.cfg, i.e.,
[email]
email_backend = airflow.contrib.utils.sendgrid.send_email
2. configure Sendgrid specific environment variables at all Airflow instances:
SENDGRID_MAIL_FROM={your-mail-from}
SENDGRID_API_KEY={your-sendgrid-api-key}.
"""
if files is None:
files = []
mail = Mail()
from_email = kwargs.get('from_email') or os.environ.get('SENDGRID_MAIL_FROM')
from_name = kwargs.get('from_name') or os.environ.get('SENDGRID_MAIL_SENDER')
mail.from_email = Email(from_email, from_name)
mail.subject = subject
mail.mail_settings = MailSettings()
if sandbox_mode:
mail.mail_settings.sandbox_mode = SandBoxMode(enable=True)
# Add the recipient list of to emails.
personalization = Personalization()
to = get_email_address_list(to)
for to_address in to:
personalization.add_to(Email(to_address))
if cc:
cc = get_email_address_list(cc)
for cc_address in cc:
personalization.add_cc(Email(cc_address))
if bcc:
bcc = get_email_address_list(bcc)
for bcc_address in bcc:
personalization.add_bcc(Email(bcc_address))
# Add custom_args to personalization if present
pers_custom_args = kwargs.get('personalization_custom_args', None)
if isinstance(pers_custom_args, dict):
for key in pers_custom_args.keys():
personalization.add_custom_arg(CustomArg(key, pers_custom_args[key]))
mail.add_personalization(personalization)
mail.add_content(Content('text/html', html_content))
categories = kwargs.get('categories', [])
for cat in categories:
mail.add_category(Category(cat))
# Add email attachment.
for fname in files:
basename = os.path.basename(fname)
attachment = Attachment()
attachment.type = mimetypes.guess_type(basename)[0]
attachment.filename = basename
attachment.disposition = "attachment"
attachment.content_id = '<{0}>'.format(basename)
with open(fname, "rb") as f:
attachment.content = base64.b64encode(f.read()).decode('utf-8')
mail.add_attachment(attachment)
_post_sendgrid_mail(mail.get())
def test_get_email_address_single_email(self):
emails_string = '*****@*****.**'
assert get_email_address_list(emails_string) == [emails_string]
def test_get_email_address_tuple(self):
emails_tuple = ('*****@*****.**', '*****@*****.**')
self.assertEqual(get_email_address_list(emails_tuple), EMAILS)
def test_get_email_address_single_email(self):
emails_string = '*****@*****.**'
self.assertEqual(get_email_address_list(emails_string),
[emails_string])
def send_email(to,
subject,
html_content,
files=None,
cc=None,
bcc=None,
sandbox_mode=False,
**kwargs):
"""
Send an email with html content using `Sendgrid <https://sendgrid.com/>`__.
To use this plugin:
0. Include ``sendgrid`` subpackage as part of your Airflow installation, e.g.,::
pip install 'apache-airflow[sendgrid]'
1. Update ``email_backend`` property in `[email]`` section in ``airflow.cfg``, i.e.,::
[email]
email_backend = airflow.contrib.utils.sendgrid.send_email
2. Configure Sendgrid specific environment variables at all Airflow instances:::
SENDGRID_MAIL_FROM={your-mail-from}
SENDGRID_API_KEY={your-sendgrid-api-key}.
"""
if files is None:
files = []
mail = Mail()
from_email = kwargs.get('from_email') or os.environ.get(
'SENDGRID_MAIL_FROM')
from_name = kwargs.get('from_name') or os.environ.get(
'SENDGRID_MAIL_SENDER')
mail.from_email = Email(from_email, from_name)
mail.subject = subject
mail.mail_settings = MailSettings()
if sandbox_mode:
mail.mail_settings.sandbox_mode = SandBoxMode(enable=True)
# Add the recipient list of to emails.
personalization = Personalization()
to = get_email_address_list(to)
for to_address in to:
personalization.add_to(Email(to_address))
if cc:
cc = get_email_address_list(cc)
for cc_address in cc:
personalization.add_cc(Email(cc_address))
if bcc:
bcc = get_email_address_list(bcc)
for bcc_address in bcc:
personalization.add_bcc(Email(bcc_address))
# Add custom_args to personalization if present
pers_custom_args = kwargs.get('personalization_custom_args', None)
if isinstance(pers_custom_args, dict):
for key in pers_custom_args.keys():
personalization.add_custom_arg(
CustomArg(key, pers_custom_args[key]))
mail.add_personalization(personalization)
mail.add_content(Content('text/html', html_content))
categories = kwargs.get('categories', [])
for cat in categories:
mail.add_category(Category(cat))
# Add email attachment.
for fname in files:
basename = os.path.basename(fname)
with open(fname, "rb") as file:
content = base64.b64encode(file.read()).decode('utf-8')
attachment = Attachment(file_content=content,
file_type=mimetypes.guess_type(basename)[0],
file_name=basename,
disposition="attachment",
content_id=f"<{basename}>")
mail.add_attachment(attachment)
_post_sendgrid_mail(mail.get())
def test_get_email_address_list(self):
emails_list = ['*****@*****.**', '*****@*****.**']
self.assertEqual(
get_email_address_list(emails_list), EMAILS)
def send_email( # pylint: disable=too-many-locals
to: AddressesType,
subject: str,
html_content: str,
files: Optional[AddressesType] = None,
cc: Optional[AddressesType] = None,
bcc: Optional[AddressesType] = None,
sandbox_mode: bool = False,
conn_id: str = "sendgrid_default",
**kwargs,
) -> None:
"""
Send an email with html content using `Sendgrid <https://sendgrid.com/>`__.
.. note::
For more information, see :ref:`email-configuration-sendgrid`
"""
if files is None:
files = []
mail = Mail()
from_email = kwargs.get('from_email') or os.environ.get(
'SENDGRID_MAIL_FROM')
from_name = kwargs.get('from_name') or os.environ.get(
'SENDGRID_MAIL_SENDER')
mail.from_email = Email(from_email, from_name)
mail.subject = subject
mail.mail_settings = MailSettings()
if sandbox_mode:
mail.mail_settings.sandbox_mode = SandBoxMode(enable=True)
# Add the recipient list of to emails.
personalization = Personalization()
to = get_email_address_list(to)
for to_address in to:
personalization.add_to(Email(to_address))
if cc:
cc = get_email_address_list(cc)
for cc_address in cc:
personalization.add_cc(Email(cc_address))
if bcc:
bcc = get_email_address_list(bcc)
for bcc_address in bcc:
personalization.add_bcc(Email(bcc_address))
# Add custom_args to personalization if present
pers_custom_args = kwargs.get('personalization_custom_args')
if isinstance(pers_custom_args, dict):
for key in pers_custom_args.keys():
personalization.add_custom_arg(
CustomArg(key, pers_custom_args[key]))
mail.add_personalization(personalization)
mail.add_content(Content('text/html', html_content))
categories = kwargs.get('categories', [])
for cat in categories:
mail.add_category(Category(cat))
# Add email attachment.
for fname in files:
basename = os.path.basename(fname)
with open(fname, "rb") as file:
content = base64.b64encode(file.read()).decode('utf-8')
attachment = Attachment(
file_content=content,
file_type=mimetypes.guess_type(basename)[0],
file_name=basename,
disposition="attachment",
content_id=f"<{basename}>",
)
mail.add_attachment(attachment)
_post_sendgrid_mail(mail.get(), conn_id)
def test_get_email_address_colon_sep_string(self):
emails_string = '[email protected]; [email protected]'
assert get_email_address_list(emails_string) == EMAILS
def test_get_email_address_colon_sep_string(self):
emails_string = '[email protected]; [email protected]'
self.assertEqual(
get_email_address_list(emails_string), EMAILS)
def test_get_email_address_list(self):
emails_list = ['*****@*****.**', '*****@*****.**']
assert get_email_address_list(emails_list) == EMAILS
def manage_slas(self, dag: DAG, session: Session = None) -> None:
"""
Finding all tasks that have SLAs defined, and sending alert emails
where needed. New SLA misses are also recorded in the database.
We are assuming that the scheduler runs often, so we only check for
tasks that should have succeeded in the past hour.
"""
self.log.info("Running SLA Checks for %s", dag.dag_id)
if not any(isinstance(ti.sla, timedelta) for ti in dag.tasks):
self.log.info(
"Skipping SLA check for %s because no tasks in DAG have SLAs",
dag)
return
qry = (session.query(
TI.task_id,
func.max(TI.execution_date).label('max_ti')).with_hint(
TI, 'USE INDEX (PRIMARY)',
dialect_name='mysql').filter(TI.dag_id == dag.dag_id).filter(
or_(TI.state == State.SUCCESS,
TI.state == State.SKIPPED)).filter(
TI.task_id.in_(dag.task_ids)).group_by(
TI.task_id).subquery('sq'))
max_tis: List[TI] = (session.query(TI).filter(
TI.dag_id == dag.dag_id,
TI.task_id == qry.c.task_id,
TI.execution_date == qry.c.max_ti,
).all())
ts = timezone.utcnow()
for ti in max_tis:
task = dag.get_task(ti.task_id)
if task.sla and not isinstance(task.sla, timedelta):
raise TypeError(
f"SLA is expected to be timedelta object, got "
f"{type(task.sla)} in {task.dag_id}:{task.task_id}")
dttm = dag.following_schedule(ti.execution_date)
while dttm < timezone.utcnow():
following_schedule = dag.following_schedule(dttm)
if following_schedule + task.sla < timezone.utcnow():
session.merge(
SlaMiss(task_id=ti.task_id,
dag_id=ti.dag_id,
execution_date=dttm,
timestamp=ts))
dttm = dag.following_schedule(dttm)
session.commit()
# pylint: disable=singleton-comparison
slas: List[SlaMiss] = (
session.query(SlaMiss).filter(SlaMiss.notification_sent == False,
SlaMiss.dag_id == dag.dag_id) # noqa
.all())
# pylint: enable=singleton-comparison
if slas: # pylint: disable=too-many-nested-blocks
sla_dates: List[datetime.datetime] = [
sla.execution_date for sla in slas
]
fetched_tis: List[TI] = (session.query(TI).filter(
TI.state != State.SUCCESS, TI.execution_date.in_(sla_dates),
TI.dag_id == dag.dag_id).all())
blocking_tis: List[TI] = []
for ti in fetched_tis:
if ti.task_id in dag.task_ids:
ti.task = dag.get_task(ti.task_id)
blocking_tis.append(ti)
else:
session.delete(ti)
session.commit()
task_list = "\n".join(sla.task_id + ' on ' +
sla.execution_date.isoformat()
for sla in slas)
blocking_task_list = "\n".join(ti.task_id + ' on ' +
ti.execution_date.isoformat()
for ti in blocking_tis)
# Track whether email or any alert notification sent
# We consider email or the alert callback as notifications
email_sent = False
notification_sent = False
if dag.sla_miss_callback:
# Execute the alert callback
self.log.info('Calling SLA miss callback')
try:
dag.sla_miss_callback(dag, task_list, blocking_task_list,
slas, blocking_tis)
notification_sent = True
except Exception: # pylint: disable=broad-except
self.log.exception(
"Could not call sla_miss_callback for DAG %s",
dag.dag_id)
email_content = f"""\
Here's a list of tasks that missed their SLAs:
<pre><code>{task_list}\n<code></pre>
Blocking tasks:
<pre><code>{blocking_task_list}<code></pre>
Airflow Webserver URL: {conf.get(section='webserver', key='base_url')}
"""
tasks_missed_sla = []
for sla in slas:
try:
task = dag.get_task(sla.task_id)
except TaskNotFound:
# task already deleted from DAG, skip it
self.log.warning(
"Task %s doesn't exist in DAG anymore, skipping SLA miss notification.",
sla.task_id)
continue
tasks_missed_sla.append(task)
emails: Set[str] = set()
for task in tasks_missed_sla:
if task.email:
if isinstance(task.email, str):
emails |= set(get_email_address_list(task.email))
elif isinstance(task.email, (list, tuple)):
emails |= set(task.email)
if emails:
try:
send_email(emails,
f"[airflow] SLA miss on DAG={dag.dag_id}",
email_content)
email_sent = True
notification_sent = True
except Exception: # pylint: disable=broad-except
Stats.incr('sla_email_notification_failure')
self.log.exception(
"Could not send SLA Miss email notification for DAG %s",
dag.dag_id)
# If we sent any notification, update the sla_miss table
if notification_sent:
for sla in slas:
sla.email_sent = email_sent
sla.notification_sent = True
session.merge(sla)
session.commit()
def test_get_email_address_tuple(self):
emails_tuple = ('*****@*****.**', '*****@*****.**')
assert get_email_address_list(emails_tuple) == EMAILS
