Hi,
I have a question about using readiness/liveness probes for services that do not expose APIs but instead process data, such as reading from a queue, processing it, and passing it to another service.
In my case, such a service can be configured by a user. Incorrect configurations occur occasionally, and users can change the configuration live without requiring a restart, as the service will reconfigure itself automatically. However, if the service is incorrectly configured, I need to be aware of it. Would using a readiness probe in this context be inappropriate? The probe could report that the service is not ready.
Most examples and documentation explain probes in the context of services that expose APIs, where Kubernetes does not route traffic to services that are not ready.
Marcin