add-alert-policy

Configure alert escalation policies by defining ordered, multi-step delivery chains with per-step delays. The policy resolves on-call rotation recipients deterministically using round-robin logic and queues delayed steps for execution.

Route system alerts by applying a severity_min threshold to your notification policy. This ensures low-severity events are skipped entirely, triggering escalation chains only for critical operational health events.

Add delayed notification steps by scheduling ordered delivery chains within the alert policy. You can configure an immediate email, an SMS after a 15-minute delay, and route to an active on-call member after 30 minutes.

Yes, the alert policy supports template-driven notifications. It uses NotificationTemplate records and renders messages with Jinja2 variables to ensure consistent content delivery across all escalation steps.

Yes, you can audit deliveries because the policy creates AlertNotificationLog entries during execution. These logs tie notifications to templates and policy steps, providing an operational audit trail for traceability.

Round-robin logic for on-call rotations resolves step recipients deterministically by evaluating priority and last_used_at timestamps. This ensures fair distribution of alerts among qualified team members.