code0Vörr AIContactGuides
Book a Demo
Guides/Shankh/Reference

Troubleshooting

Reference

An alert source is connected but no incident appears

  1. Confirm the source is sending an alert.
  2. Confirm the expected workspace is configured.
  3. Confirm the alert severity is mapped correctly.
  4. Confirm the webhook/API key is still valid, if using direct ingest.
  5. Confirm the corresponding alert rule is enabled.

An incident was raised but nobody was notified

  1. Confirm the channel in escalation step 1 is currently supported for delivery and has been tested.
  2. Confirm the on-call calendar covers the time of the incident for that workspace.
  3. Confirm the level points at a team that has members.
  4. Confirm those members have completed their channel setup.
  5. Move the step to a channel that is delivering if it still fails.

A webhook stopped working after changing its key

  1. Regenerating the API key invalidates the old key immediately.
  2. Update every sender with the new key.
  3. Send a test request and confirm a new incident appears.

A member cannot change configuration

  1. Check the user's workspace role.
  2. Members can acknowledge and resolve incidents, but administrative configuration requires Admin access.
  3. Promote the user to Admin only if they genuinely need to change configuration.

The mobile app is not receiving alerts

  1. Confirm the app is connected to the correct Shankh server.
  2. Confirm the user is signed in.
  3. Confirm notification permission is enabled on the device.
  4. Confirm the device has registered for Shankh push notifications.
  5. Confirm the user's current product access includes mobile push.

An incident has no AI analysis

  1. Expected — AI incident analysis is not implemented in Shankh.
  2. There is no provider picker, key field or model choice to configure.
  3. The AI summary and AI postmortem in Reports are text templates built from the incident record.
  4. Alert delivery, escalation, acknowledgement and resolution are unaffected.

Metrics look wrong or incomplete

  1. Confirm incidents are being resolved rather than left active.
  2. MTTA and MTTR are calculated from resolved incidents only.
  3. Encourage the acknowledge → investigate → resolve sequence.
  4. Re-check the report once the open incidents are closed.

Something else, or you need support

  1. Shankh does not currently have a dedicated public support channel.
  2. For general product support, contact hello@frigga.cloud.