Multi-cluster-observability

This document show how to configure and install multi-cluster obsavability solution using Grafana, Prometheus, Thanos, and Loki.

Background

• We want to set up a central observability solution for all of three clusters in AWS VPC.
• We assume the three clusters' names are development, pre-production, and production.
• The central observer will be installed in the development cluster. It will also monitor the development cluster itself.

Prerequisites

• IAM with creating role, policy privilege
• Helm
• Helm chart of kube-prometheus, available at Link
• Helm chart of thanos, available at Link
• Helm chart of envoy proxy, available at Link
• Helm chart of Grafana-loki
• EKS
• kubectrl

Steps

Following the steps below:

• Individual cluster config and installation (Loki will be updated later.)

• Observer cluster config and installation (Loki will be updated later.)