We currently use the health checks of axon and also of our services.
But unfortunately we also have the situation that the basic health checks are green but we see NoHandlerForCommandException because the services are not connected properly. And even in the dashboards the service is visible.
I will try the heartbeat but still if would be great if we would be able that the health check fails if a service is not connected properly to the server.
Are there already metrics we can check? The NoHandlerForCommandException is a warning only. Can thiis also be exposed to Prometheus?
unfortunately, from a generic perspective, Axon can’t assume that a “NoHandlerForCommandException” is an indicator of an unhealthy application. AxonServer does expose information about which applications expose which handlers in its APIs. Check /swagger-ui.html on your AxonServer for an overview of all the calls you can make.
We do have plans for exposing more APIs on the AxonServer Connector to fetch information from AxonServer. This would be to allow replays to be triggered by your applications. Adding an option to retrieve certain groups of command handlers could be a nice addition too. You could then use that to implement an application-specific health check.
Right.The NoHandlerForCommandException is only a indicator for us as we know it should be there. We running our services in different environments where this is “normal” as this service does not exist.