Create monitor dependencies

While Datadog does fully support composite monitors, there is not an official way to create alerting trees.

Some Datadog users have combined webhook notifications with downtime scoping via the Datadog API to achieve a similar result.

At a high level, the setup for this is as follows:

  • Alert A triggers and has an @webhook-notification.
  • Notification reaches out to the Datadog downtime API by $scope to mute any other alerts.
  • When Alert A resolves, use a different @webhook-notification to remove the downtimes from the same $scope. It should be noted that this can impact previously scheduled downtimes if you have an active downtime overlapping with the defined $scope.

First, create the webhooks:


Full text for API endpoints (2nd input box for each in the left column):



And the webhook content for both:

{"scope": "$ALERT_SCOPE"}

Then, create “Alert A” - for example- a no-data alert for a grouped percentage of hosts for each Availability zone.


Then, in the alert message, you’ll want to use the @notify webhook to mute all subsequent hosts in that Availability Zone when it triggers, and unmute when the alert resolves:


Here is that full sample markup:

That's alot of missing data - check first to see if there is an AWS outage?
{{}} is missing 50% of data!! ALL OTHER ALERTS FOR {{}} WILL BE AUTOMUTED

{{}} is NO LONGER missing 50% of data!! ALL OTHER ALERTS FOR {{}} ARE UNMUTED