Skip to main content

I am new to pagerduty. Trying to set up alerts and escalation policy. nothing seems to work at all. 

I have set escalation policy repeat alert 9 times if no one acknowledges. and guest what after that change I stop getting any kind of alerts even if I acknowledged it whilst alert is still firing from grafana side. 

 

I doesnt repeat any alert at all for some reason and also creates incidents in a single payload. 

Why would I want all incidents in one payload. 

Also do you have feature when alert closes itself once it stops firing from grafana side like it happens in opsgenie. 

Hi ​@msaeed! Welcome to PagerDuty, we’re happy to help get this behaving like you expect.

First, let me highlight a couple of things to clarify how PD works compared to Opsgenie:

  • Escalation policy “repeat X times” only applies while the incident is still unacknowledged. If someone acknowledges, PD won’t continue escalating or repeating, and that’s by design. If you want acknowledged incidents to re-notify after a period, enable Acknowledgement Timeout on the service to “re-trigger acknowledged incidents” after N minutes and re-notify the on-call. Docs: Escalation Policy Basics, Configurable Service Settings → Acknowledgement Timeout.

  • If it felt like alerts “stopped” after your change, it’s often because the incident was already acknowledged (so PD stops paging), the service urgency is low (low-urgency incidents don’t escalate), or alert grouping is enabled (new alerts are being grouped into the open incident to reduce noise). Quick checks:

  • “All incidents in one payload” sounds like Alert Grouping. You can tune or turn that off so new alerts create new incidents instead of grouping into one.

  • Auto-close when Grafana stops firing: Grafana’s PagerDuty contact point sends a resolve when the alert returns to OK (unless disabled). If your tool doesn’t send resolves consistently, you can also enable Auto-Resolution on the PagerDuty service to auto-close incidents after a set time. Grafana docs: Configure PagerDuty for Alerting. PD Auto-Resolution: Configurable Service Settings.

  • Deduplication note for Grafana: PD will dedup by the Events API v2 dedup_key. Grafana typically generates this automatically; controlling it isn’t always exposed, which can make “always update the same incident” tricky. Some background here: Grafana forum: dedup key behavior. If you want fewer incidents/noises, consider PD Alert Grouping or ensure your Grafana alerts keep a stable identity for dedup.

Also worth a quick health check:

  • Make sure someone is on call across your policy levels; if there are gaps, incidents may not trigger at all. Community best practices.
  • If you still aren’t receiving pages as expected, this guide lists common reasons (support hours, urgency, suppression, etc.): Expected Notification Behavior.

If you’re still struggling to get your setup up and running, I’d recommend reaching out to Customer Support. They’ll be happy to look into your account and help you get the config you’re looking for. Please make sure to use this topic as reference!

Hope this helps, and enjoy your day!

Terms & ConditionsCookie settingsAccessibility statement