Objective: Learn how to configure outlier detection in Envoy and how to visualize this pattern and its affects in Grafana.
First, navigate to this exercise’s directory:
cd ~/service-mesh-training/exercises/lab-04/04-outlier-detection/
Take a look at the updated Kubernetes config files in files/app and see where we are implementing outlier detection.
Now apply the updated config files:
kubectl apply -f files/app
Look at the dashboard, now you can see that Envoy is ejecting the broken service from the load balancing.
The service time has also reduced as the retry is not as spammy.