#database-alerts.
Set it up
1
Have a source and a destination
You need an incoming integration that produces the alerts (CloudWatch, Grafana, Sentry, the Incoming API, and so on) and at least one place to send them: a Slack channel or an outgoing webhook. Both are configured on the team or one of its services. See Slack and Outgoing Webhooks.
2
Create the rule
Open the team, go to On-Call, expand Alert Rules, and select Add Rule.
- Under Integration scope, pick the integration the alerts come from. Leave it empty to match alerts from any integration.
- Add conditions that describe the alerts you want. Match on the alert’s name, severity, tags, or any of its details.
- Set External notifications to Send only to selected destinations and pick your Slack channels or webhooks.
- Set On-call paging to Do not page anyone if these alerts should not page.
3
Check it
Trigger a matching alert from the source. It shows up under Alerts → Routed with the rule that matched it and where it was sent.
Example
Send production RDS alarms to Slack and a webhook, without paging:
Everything else from that integration keeps flowing to the rules below it and, if nothing matches, to the team’s normal paging.
Notifications and paging are separate choices
Every rule makes two decisions on its own:- External notifications: use the team’s usual destinations, send only to the ones you picked, or send nothing.
- On-call paging: page as usual, or do not page anyone.
Routes apply when an alert fires or fires again. Acknowledging or resolving an alert in Warrn is not forwarded to the destination.
Where routed alerts show up
Alerts that were routed without paging live in Alerts → Routed. They are kept out of the Active and Open queues so those stay focused on what needs a person, but they are always in All Alerts and search. Each one shows whether the source is still firing and records that it was routed without paging.Good to know
- Rules run top to bottom, first match wins. Put specific routes above broad ones.
- Destination severity filters still apply. If a Slack channel is set to only receive high and critical, a medium alert routed to it is skipped.
- A failed delivery does not fall back to paging. Check the destination’s Observability tab to see recent deliveries and any errors.