Point a Grafana alerting contact point at OpsPing and every rule transition pages on-call: alertname becomes the message, severity maps to P1–P5, the fingerprint becomes a stable alias, and the Grafana externalURL rides into the alert details so the page links back to the panel.
Contact point setup
- Create the OpsPing integration (Grafana source type) with a default service, copy the webhook URL.
- Grafana → Alerting → Contact points → add a Webhook contact point with that URL.
- Attach the contact point to your notification policies; enable resolve messages so resolutions close alerts.
- Fire a test rule and verify priority, service icon, and the details link.
Field mapping
| Grafana webhook field | OpsPing alert |
|---|---|
labels.alertname | Message |
labels.severity | Priority — P1–P5 |
| Alert fingerprint | Alias — stable, dedupes repeat firings |
externalURL | Details — the page links back to the Grafana panel |
| Resolved entries | Close the matching alert |
Grafana's alerting webhook sends the same { "alerts": [...] } shape as Alertmanager, so multi-alert notifications are processed entry by entry with a per-entry tally in the response.
Coming from Grafana OnCall?
With OnCall OSS archived (March 2026), many teams keep Grafana for alerting and move only the paging layer — exactly this integration. The Grafana OnCall migration guide covers retiring Cloud Connection and rebuilding schedules and escalation chains on OpsPing, and the full comparison covers when Grafana Cloud IRM is the better call instead.
- Route noise with alert policies before it pages — match on source, tags, or severity.
- One integration per Grafana org (or per service) keeps default-service icons and team scoping clean.
FAQ
Does this work with Grafana OSS and Grafana Cloud?
Yes — both emit the same alerting webhook shape. OpsPing auto-detects it; the only Grafana-side difference is where the contact point lives (Alerting → Contact points in both).
How do I stop Grafana test notifications from paging everyone?
Send tests to a dedicated integration whose default service is a sandbox, or add an alert policy that suppresses that source/tag combination. Suppressed alerts are stored and visible — they just don't notify.
Do resolved Grafana alerts close the OpsPing alert?
Yes, when resolve messages are enabled on the notification policy: the resolved entry closes whichever alert owns that fingerprint's alias, even if it was already acknowledged.