-
Notifications
You must be signed in to change notification settings - Fork 18
Monitoring
Ratel provides a few monitoring capabilities out of the box:
Services renew their registration at registry server every 60 seconds, so that when the service provider is down or looses connection to the registry server, its endpoints are removed from the registry and this fact is propagated to all clients.
In other words registry server only contains entries for 'alive' services.
Ratel can log execution time on both client and server time, so that you can observe any anomalies appearing in this. This feature combined with Process tracing allows you to measure execution times and communication lags along remote service calls.
Every service registered in Ratel can be sent a heartbeat request. The request is sent to a service url (which can be retrieved from Ratel) with a http HEAD method. The default handler will return http 200 response. There is an issue that will allow service to implement custom healthcheck handler Issue #41. The MonitoringApplication is an example how to use heartbeats to check services aviability.
Each communication between Ratel client and Ratel-exposed-service triggers events on both client side. More on this topic in the Monitoring events page.