docker-screammachine

A docker image that alerts in slack if something fishy is happening in streammachine.

It does a request to streammachine's /api/streams endpoint and looks at the last_ts attribute of each stream. If that timestamp is more than SECONDS_OFFSET behind, it posts a message to a slack webhook.

If you've provided a streams.json in the same directory as index.json or mount it as a volume into your container, this script will also look to see if there is a stream that it can recreate from it. If it can, it will delete that existing stream from primary and recreate it.

Screammachine can also alert by checking the streamdash api URL and finding the listeners count. If the listeners count fall below a certain threshold, it will alert (see optional env variables below).

requirements

• slack webhook url (https://api.slack.com/incoming-webhooks)
• node v10.16.3

deployment

The following environment variables need to be provided for this to work:

• SOURCE_URL the /api/sources endpoint in StreamMachine. (Example: http://streammachine_url/api/streams) If this endpoint is in a private network, then this script needs to either run within that network or have access to that endpoing.
• SECONDS_OFFSET the threshold of number of seconds offset you'll tolerate before you want this script to alert. (Example: 30)
• WEBHOOK_URL slack webhook URL. (Example: https://hooks.slack.com/services/LISAFRANK/BLAHBLAHUNICORNS124)
• PREPEND_MESSAGE what you'd like to prepend to the slack message. For example, if you want to @channel here or add some emoji's like :hear_no_evil: :radio:

optional

• LOW_LISTENERS_THRESHOLD
• STREAMDASH_HOURS_URL the URL that your streamdash's hours api URL is at. Example: http://streamdash.com/api/hour

This can be set up to run as a regular cronjob on a machine (node index.js), or set up as a container running as a kubernetes cronjob.

development

The meat of the script is in index.js