Skip to main content

Notification Health Reference

Summary​

One health route exists. Health checks are registered as a service but no individual check is added, so the endpoint reports on an empty check set.

Audience​

DevOps engineers, engineers, support, QA and architects.

Reference Content​

The verified reference material for this topic is set out in the sections below.

The health route​

Mapped at a single path and tagged for health. It declares no authentication and no authorization. Because the check collection is empty, it returns the framework default response — a plain overall status — rather than a structured per-dependency report.

What a healthy response proves​

QuestionAnswered by the health endpoint
Is the HTTP host accepting requestsYes
Is the process aliveYes
Is PostgreSQL reachableNo
Is the broker connectedNo
Is the consumer subscribedNo
Is any channel provider workingNo
Is the failed queue growingNo

The gap between the first two rows and the rest is the central operational risk. A service whose broker connection has been down for hours, consuming nothing, returns a healthy response throughout.

Because startup fails when the database is unreachable, a responding service does imply the database was reachable at startup — but not that it still is.

Service information​

A second operational route returns service identity: name, description, resolved environment name and the current instant, wrapped in the platform response envelope with the request's correlation identifier.

It performs no dependency probing and is not a health check. It is useful for confirming which service and environment answered, and for verifying that correlation propagation is working end to end.

The environment name is the only environment detail disclosed; no host, connection or configuration value is returned.

Probe guidance​

Recorded as an observation from source rather than a recommendation: because the endpoint covers only HTTP liveness, using it as a readiness probe would mark an instance ready before the consumer has connected, and would keep it ready while consumption is broken.

The only positive signal that consumption is active is the consumer's bound-to-queue log line. See logging reference.

Verified absence​

There is no separate readiness route, no liveness route, no tag-filtered health route, no health-check UI, no database check, no broker check, no hosted-service liveness check and no startup probe.

Classification​

Partial — the endpoint is Implemented; the check set is empty.

Requires confirmation​

Which endpoint orchestration probes consume today, and whether dependency checks are planned, require confirmation.

See Also​

Keywords​

  • Notification database
  • Health Reference
  • Draft database documentation

Source References​

  • microservices/src/notification-service/Program.cs
  • microservices/src/shared-kernel/Extensions/EndpointRouteBuilderExtensions.cs
  • microservices/src/notification-service/Messaging/NotificationEventConsumer.cs

Revision Information​

  • Status: Draft
  • Last reviewed: 2026-07-21
  • Review cycle: Quarterly