Skip to content

Instantly share code, notes, and snippets.

@devdazed
Last active September 10, 2024 20:00
Show Gist options
  • Save devdazed/473ab227c323fb01838f to your computer and use it in GitHub Desktop.
Save devdazed/473ab227c323fb01838f to your computer and use it in GitHub Desktop.
Updates a Slack User Group with People that are on call in PagerDuty
#!/usr/bin/env python
from __future__ import print_function
import json
import logging
from urllib2 import Request, urlopen, URLError, HTTPError
from base64 import b64decode
log = logging.getLogger(__name__)
class SlackOnCall(object):
# The Slack API token to use for authentication to the Slack WebAPI
slack_token = None
# The Pager Duty API token to use for authentication into the PagerDuty API
pager_duty_token = None
# The domain prefix for the PagerDuty installation (acme.pagerduty.com will be "acme")
pager_duty_domain_prefix = None
# The Slack @user-group to update (Default: oncall)
slack_user_group_handle = 'oncall'
# The maximum escalation level to add to the group
# (eg. if escalation level = 2, then levels 1 and 2 will be a part of the group
# but not any levels 3 and above.
escalation_level = 2
def __init__(self, slack_token, pager_duty_token, pager_duty_domain_prefix,
slack_user_group_handle=slack_user_group_handle, log_level='INFO',
escalation_level=escalation_level):
self.slack_token = slack_token
self.pager_duty_token = pager_duty_token
self.pager_duty_domain_prefix = pager_duty_domain_prefix
self.slack_user_group_handle = slack_user_group_handle
self.escalation_level = escalation_level
self._slack_user_group = None
self._on_call_email_addresses = None
self._all_slack_users = None
log.setLevel(log_level)
def run(self):
"""
Gets user group information and on-call information then updates the
on-call user group in slack to be the on-call users for escalation
levels 1 and 2.
"""
slack_users = self.slack_users_by_email(self.on_call_email_addresses)
if not slack_users:
log.warning('No Slack users found for email addresses: %s', ','.join(self.on_call_email_addresses))
return
slack_user_ids = [u['id'] for u in slack_users]
if set(slack_user_ids) == set(self.slack_user_group['users']):
log.info('User group %s already set to %s', self.slack_user_group_handle, slack_user_ids)
return
self.update_on_call(slack_users)
log.info('Job Complete')
@staticmethod
def _make_request(url, body=None, headers={}):
req = Request(url, body, headers)
log.info('Making request to %s', url)
try:
response = urlopen(req)
body = response.read()
try:
return json.loads(body)
except ValueError:
return body
except HTTPError as e:
log.error("Request failed: %d %s", e.code, e.reason)
except URLError as e:
log.error("Server connection failed: %s", e.reason)
@property
def slack_user_group(self):
"""
:return: the Slack user group matching the slack_user_group_handle
specified in the configuration
"""
if self._slack_user_group is not None:
return self._slack_user_group
url = 'https://slack.com/api/usergroups.list?token={}&include_users=1'.format(self.slack_token)
groups = self._make_request(url)['usergroups']
for group in groups:
if group['handle'] == self.slack_user_group_handle:
self._slack_user_group = group
return group
raise ValueError('No user groups found that match {}'.format(self.slack_user_group_handle))
@property
def on_call_email_addresses(self):
"""
Hits the PagerDuty API and gets level 1 and level 2 escalation
on-call users and returns their email addresses
:return: All on-call email addresses within the escalation bounds
"""
if self._on_call_email_addresses is not None:
return self._on_call_email_addresses
url = 'https://{}.pagerduty.com/api/v1/users/on_call'.format(self.pager_duty_domain_prefix)
on_call = self._make_request(url, headers={'Authorization': 'Token token=' + self.pager_duty_token})
users = set() # users can be in multiple schedule, this will de-dupe
for user in on_call['users']:
for schedule in user['on_call']:
if schedule['level'] <= self.escalation_level:
users.add(user['email'])
log.info('Found %d users on-call', len(users))
self._on_call_email_addresses = users
return users
@property
def all_slack_users(self):
if self._all_slack_users is not None:
return self._all_slack_users
url = 'https://slack.com/api/users.list?token={}'.format(self.slack_token)
users = self._make_request(url)['members']
log.info('Found %d total Slack users', len(users))
self._all_slack_users = users
return users
def slack_users_by_email(self, emails):
"""
Finds all slack users by their email address
:param emails: List of email address to find users
:return: List of Slack user objects found in :emails:
"""
users = []
for user in self.all_slack_users:
if user['profile'].get('email') in emails:
users.append(user)
return users
def update_on_call(self, slack_users):
"""
Updates the specified user-group
:param slack_users: Slack users to modify the group with
"""
user_ids = [u['id'] for u in slack_users]
url = 'https://slack.com/api/usergroups.users.update?token={0}&usergroup={1}&users={2}'.format(
self.slack_token,
self.slack_user_group['id'],
','.join(user_ids)
)
log.info('Updating user group %s from %s to %s',
self.slack_user_group_handle, self.slack_user_group['users'], user_ids)
self._make_request(url)
def lambda_handler(*_):
"""
Main entry point for AWS Lambda.
Variables can not be passed in to AWS Lambda, the configuration
parameters below are encrypted using AWS IAM Keys.
"""
# Boto is always available in AWS lambda, but may not be available in
# standalone mode
import boto3
# To generate the encrypted values, go to AWS IAM Keys and Generate a key
# Then grant decryption using the key to the IAM Role used for your lambda
# function.
#
# Use the command `aws kms encrypt --key-id alias/<key-alias> --plaintext <value-to-encrypt>
# Put the encrypted value in the configuration dictionary below
encrypted_config = {
'slack_token': '<ENCRYPTED VALUE>',
'pager_duty_token': '<ENCRYPTED VALUE>'
'pager_duty_domain_prefix': '<ENCRYPTED VALUE>'
}
kms = boto3.client('kms')
config = {x: kms.decrypt(CiphertextBlob=b64decode(y))['Plaintext'] for x, y in encrypted_config.iteritems()}
return SlackOnCall(**config).run()
def main():
"""
Runs the Slack PagerDuty OnCall group updater as a standalone script
"""
from argparse import ArgumentParser
parser = ArgumentParser(usage=main.__doc__)
parser.add_argument('-st', '--slack-token', required=True, dest='slack_token',
help='Slack token to use for auth into the Slack WebAPI')
parser.add_argument('-su', '--slack-user-group', dest='slack_user_group_handle', default='oncall',
help='Slack user group to add on-call users to. (Default: oncall)')
parser.add_argument('-pt', '--pager-duty-token', required=True, dest='pager_duty_token',
help='PagerDuty token to use for auth into the PagerDuty API')
parser.add_argument('-pd', '--pager-duty-domain-prefix', required=True, dest='pager_duty_domain_prefix',
help='Your domain prefix for PagerDuty')
parser.add_argument('-el', '--max-escalation-level', dest='escalation_level', default=2, type=int,
help='Max escalation level to add on-call users for group. (Default: 2)')
logging.basicConfig()
args = vars(parser.parse_args())
SlackOnCall(**args).run()
if __name__ == '__main__':
main()
@devdazed
Copy link
Author

PagerDuty Slack OnCall Group

Use PagerDuty and Slack? Create a Slack @oncall group containing the people that are currently in specified escalation levels in PagerDuty. Just create a Slack user group called oncall then run this script.

NOTE: Users are matched by email address, the email address they use in Slack must match the one they use in PagerDuty.

./slack-pagerduty-oncall.py --slack-token 'xoxp-xxxxxxxx-xxxxxxx-xxxxxxx' \
                                            --pager-duty-token 'abcdefghijklmnopqrstuvwxyz' \
                                            --pager-duty-domain-prefix 'acme'

Running as a Cron job

Just ignore, or remove, the lambda_handler function and run the above example as a scheduled cron job. Can be set to the dates/times your PagerDuty schedules change.

Running as an AWS Lambda method

Use AWS IAM keys to encrypt the configuration in the lambda_handler method and set up your AWS Lambda method to run on a schedule. The schedule should be set, just as in the cron setup, to coincide with your PagerDuty schedule changes.

Copy link

ghost commented May 24, 2017

awesome script. Was definitely looking for something like this.

@markddavidoff
Copy link

I've updated this to work with pagerduty v2 api:
https://gist.github.com/markddavidoff/863d77c351672345afa3fd465a970ad6

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment