We want the service mesh to be observable.
Kubernetes, Calico (the network data plane for our service mesh), Kuma can send metrics to Prometheus:
- https://docs.projectcalico.org/reference/felix/prometheus
- https://kuma.io/docs/0.7.1/policies/traffic-metrics/#secure-dataplane-metrics
- https://kubernetes.io/docs/tasks/debug-application-cluster/resource-usage-monitoring/
Plan is to:
- deploy Prometheus on Kubernetes cluster
- enable Kuma metrics policy
- send Calico metrics
- send Kubernetes metrics
- publish Grafana dashboards