I am new to this Istio world and I have been asked to implement the Istio Circuit Breaking and outlier detection in one of our service mesh where we have Istio in place with default numbers running on it ( ** we are using Istio as open source) , I would really appreciate if somebody could help me with the queries I have wrt Istio Circuit breaking and outlier detection concepts-
I am aware via Istio Circuit breaking implementation, we get 5xx errors which captures latency details. So the question is - Since we have readiness probe in place as well for health check validation of pods, which action get performed first - is it readiness probe or Istio configurations when traffic is routed let say at 3x load in the service mesh.
2. Since we have readiness probe in place, what is the significance of outlier detection feature in Istio? Where do they differ ? a practical example would be helpful.
3. Can the configurations of connection pool and outlier detection be set as dynamic considering autoscaling in place.
4. In my project, when I am disabling Destination Rule from the flow, the 5xx errors are coming as minimum but when we are enabling it, 5xx are increasing - is this the expected behavior ?
5. Since we have 10-15 microservices in our environment, what are the ideal practice to implement circuit breaking and outlier detection considering we need to enable for all the microservice ?
6. how to ensure which error is hit by Istio - from the output perspective, how the error looks like upstream, downstream- catchy words.. ?
7. which specific 5xx error Circuit Breaking feature targets since in the performance report we have multiple 5xx. So how to differentiate ?