Observing the service mesh

In this section, we will use Kiali together with Jaeger to observe what’s going on in the service mesh.

Before we do that, we need to understand how to get rid of some noise created by the health checks performed by Kubernetes’ liveness and readiness probes. In the previous chapters, they used the same port as the API requests. This means that Istio will collect metrics for the usage of both health checks and requests sent to the API. This will cause the graphs shown by Kiali to become unnecessarily cluttered. Kiali can filter out traffic that we are not interested in, but a simpler solution is to use a different port for the health checks.

Microservices can be configured to use a separate port for requests sent to the actuator endpoints, for example, health checks sent to the /actuator/health endpoint. The following line has been added to the common configuration file for all microservices, config-repo/application.yml:

management...