Skip to main content
Grafana OnCall OSS entered maintenance mode on 11 March 2025 and was archived on 24 March 2026. Its repository is read-only, and the Cloud-connected SMS, phone and mobile push notifications stopped working for self-hosted installs. Grafana’s recommended path is Grafana Cloud IRM; see Grafana’s OnCall OSS migration guide. Observer is not an on-call tool: it has no schedules and no escalation chains. This guide covers how to connect Observer to the tool you move paging to, and the parts Observer adds: status pages, customer updates and a measured record of each incident.

What moves where

Steps

1

Choose where paging lives

Move schedules and escalation chains to the new on-call tool. For Grafana Cloud IRM, follow Grafana’s guide linked above. A small team without schedules can page phones directly with Pushover or ntfy.
2

Connect Observer to it

In the console, open Alerts and add a destination for the tool you chose: Grafana IRM, PagerDuty, Splunk On-Call, Pushover or ntfy. Pick the Paging preset, then use Send test to confirm a test alert reaches it.
3

Give Observer the same signals

Status in Observer comes from metrics. Either point Grafana Alerting at an Observer inbound source, so the alerts you already have drive status pages (Receive alerts from your monitoring), or let the agent read your metrics directly (Define your first metric).
4

Publish a status page

Create a page, add the services your customers see, and let them subscribe. See Publish your first status page.
5

Test end to end

Make one metric unhealthy, for example by firing an alert that one of your mapping rules routes to an alert-backed metric. Check that the new on-call tool pages, that the status page shows the problem, and that the alert closes in the on-call tool when the metric recovers.
6

Retire OnCall OSS

Stop OnCall OSS only after the new on-call tool has paged someone for a test or a real alert.