Skip to main content

Helpdesk Event Retries & Failures

Outbox publisher behavior​

When the shared processor is invoked, a broker-unavailable result or exception marks the row Failed, increments RetryCount, and records an error. The Helpdesk query includes both Pending and Failed rows on every call. There is no delay, exponential backoff, maximum attempt count, lock, terminal dead-letter status or quarantine in this store.

Because Helpdesk does not schedule the processor, those retry rules are currently dormant unless an unverified external caller exists.

Consumer behavior​

Notification Service retries connection establishment after failures. An unexpected processing exception causes a negative acknowledgement with requeue. After event resolution, individual notification channel sends use the Notification Service retry policy and persist attempts; exhausted sends become failed notifications.

Error classification​

Malformed or unsupported payloads can resolve to no notification intent and be recorded as processed outcomes. Infrastructure exceptions are requeued. Operational alerting and poison-message handling for endlessly failing Helpdesk events were not found and Require confirmation.

Source References​

  • microservices/src/shared-kernel/Messaging/OutboxProcessor.cs
  • microservices/src/helpdesk-service/Infrastructure/Persistence.cs
  • microservices/src/notification-service/Messaging/NotificationEventConsumer.cs
  • microservices/src/notification-service/Application/NotificationDispatcher.cs

Revision Information​

Draft source-backed documentation reviewed on 2026-07-20; next quarterly review is 2026-10-20.