Set Up Monitoring on your k8s cluster with the Helm and Prometheus Operator#

A monitoring system usually consists of:

  • a time-series database holding metrics
  • a visualization layer
  • an alerting layer

One popular monitoring solutions is:

  • Prometheus - A time series database and monitoring tool that works by polling metrics endpoints and processing the data from them
  • Grafana - A data analytics and visualisation tool to build dashboards for your metrics
  • AlertManager - Usually deployed alongside Prometheus - deduplicating, grouping, and routing alerts

On the kubernetes side - these expose cluster and machine level metrics:

Deploying all of this can be tricky but we can make use of Helm (package manager) with the Prometheus Operator and kube-prometheus.

The prometheus-operator chart is used.

…This is a bit too advanced to dig into for now…too much complexity