A basic website with

- A Frontend
- An API
- An RDS Database
- Replicated across two availability zones for failover
- A CloudFront (CDN) to serve static assets
- A Load balancer upstreaming to the API
- May or may not be using an underlying container orchestrator

What is the effort/task that goes into observing this above Infrastructure?

• Setup Cloudwatch or Grafana Dashboards

• Setup alerting.

  • If it’s Grafana, it cannot use templatized dashboards for alerting.

• Chose among ~ 500 available exporters that match the need

  prometheus/docs

  <aside> 💡 If any of the components change, Repeat the above steps.

  </aside>

• Additionally, data is available in multiple sources. For example, an Instance has both CPU and memory metrics. Data concerning CPU is available in Cloudwatch, while Memory is usually available elsewhere via Node exporters.

  Without openmetrics-exporter

  

  With openmetrics-exporter