Inspect Delivery evidence
Trace a Notification Event through its Delivery and Delivery Attempts, then interpret production Delivery Health safely.
Start on a Notifier detail page after a real Incident opens or Recovery is confirmed. PointWarden snapshots the Verified, Enabled, subscribed Notifiers at that moment; later subscription changes do not backfill the Notification Event.
Follow the evidence chain
| Evidence | Meaning |
|---|---|
| Notification Event | An immutable record that an Incident opened or Recovery was confirmed. |
| Delivery | The at-least-once process of sending one Notification Event to one Notifier, with one stable identity across retries. |
| Delivery Attempt | One request made while trying to complete a Delivery. |
For the same Incident and Notifier, the Incident-opening Delivery becomes terminal before the Recovery Delivery begins. This ordering does not block other Monitors or Notifiers.
Inspect production and test history separately
Open Production Delivery history
Find the Incident opened or Recovery confirmed row. The row shows the Delivery status, Monitor, Attempt count, and event time.
Expand the Delivery
Read each numbered Delivery Attempt in order. Safe diagnostics include a final HTTP status when available, a failure category, duration, and attempt time.
Keep Test Delivery history separate
Test Deliveries are verification evidence. They are not Notification Events, production Deliveries, or evidence of current Delivery Health.
Interpret Delivery Health
Delivery Health is derived only from the latest production Delivery activity. It is independent of Verified or Unverified state and Enabled or Paused lifecycle.
| State | Meaning |
|---|---|
| No delivery data | No production Delivery activity exists. |
| Delivering | Initial production work is pending or leased. |
| Healthy | The latest production Delivery succeeded. |
| Retrying | A failed Delivery Attempt will be retried. |
| Failing | The latest Delivery ended permanently; future subscribed events are still attempted. |
A cancelled Delivery does not drive current Delivery Health. Pausing, deleting, or changing the Notifier can cancel unfinished work for its previous configuration.
Keep receiver delivery separate from Availability
A receiver failure can make Delivery Health Retrying or Failing, but it never changes Endpoint Availability or the Incident.
Review Incident and Recovery confirmation for the evidence that creates Notification Events. If a receiver Delivery is missing, troubleshoot a missing receiver Delivery by comparing subscriptions, verification, lifecycle, and receiver evidence. The Owner quickstart shows the setup path that precedes this evidence.
If Delivery Health is Retrying, troubleshoot Retrying Delivery Health. If Delivery Health is Failing, troubleshoot Failing Delivery Health.