Which option is considered the best practice for EKS cluster monitoring?

[Nasil114]

a) Installing Prometheus and Grafana on the EKS cluster:

• Simplifies configuration.
• Monitoring becomes inaccessible if the cluster goes down.

b) Setting up Prometheus and Grafana on external infrastructure from the EKS cluster:

• Medium complexity in implementation.
• Monitoring remains accessible if the cluster fails.

c) Deploying Prometheus with Thanos on the EKS cluster and Grafana externally:

• Thanos becomes inaccessible if the cluster fails.

d) Deploying Prometheus with a Thanos sidecar within the EKS cluster, while positioning Thanos Store, Query components, and Grafana outside the cluster:

• Ensures monitoring access even if the cluster fails.
• Requires a more complex setup.