WebMonitoring

Escalation Policies

Escalation policies define what happens when a fired alert is not acknowledged within a set time. Policies are composed of ordered levels — each level specifies who to notify and how long to wait before escalating to the next level.

Policy List

The policy list at /monitoring/escalation shows all configured policies. Each row displays: policy name, number of escalation levels, repeat setting, and the count of alert rules referencing this policy.

Creating an Escalation Policy

Click "New Policy" to launch the step-by-step policy wizard:

Step	Field	Description
1 — Basics	Name	Policy name (e.g., "Critical Ops", "FinOps Weekly")
1 — Basics	Repeat	If checked, the escalation restarts from Level 1 after reaching the final level without acknowledgment
2 — Levels	Notification Channels	Multi-select of configured channels to notify at this level
2 — Levels	Delay (minutes)	How long to wait at this level before escalating to the next level if unacknowledged
2 — Levels	On-Call Schedule	Optional. If selected, the currently on-call user is added as a notification target for this level.
3 — Review	Timeline Preview	Visual timeline showing the escalation sequence and total time to final level

Use the "+ Add Level" button to add additional escalation levels. Levels are ordered and drag-reorderable in the wizard UI.

Example Policy Configuration

Example: 3-Level Critical Ops Policy

Policy: "Critical Ops"
Repeat: true

Level 1 (immediate)
  Channels: Slack #ops-alerts
  Wait: 15 minutes

Level 2 (T+15 min if unacknowledged)
  Channels: PagerDuty, Email to on-call
  On-Call Schedule: Primary Ops Rotation
  Wait: 30 minutes

Level 3 (T+45 min if still unacknowledged)
  Channels: PagerDuty, Email to ops-manager@company.com
  On-Call Schedule: Manager Escalation
  Wait: 60 minutes (then repeat from Level 1)

Editing and Deleting Policies

Click the edit icon on any policy to open it in the wizard for modification. Deleting a policy is blocked if any alert rule currently references it. Reassign those rules to a different policy first, then delete.

Acknowledgment Halts Escalation

Acknowledging an alert (via the API, Monitoring UI, or PagerDuty webhook) immediately pauses escalation at the current level. The escalation engine will not advance to the next level once an acknowledgment is recorded, even if the delay period has not yet elapsed.