Your team wants to monitor the health and performance of the Kubernetes cluster. What exact steps would you take to set up Prometheus and Grafana for this purpose?

To monitor the health and performance of a Kubernetes cluster, I would set up Prometheus and Grafana together, since they are a powerful combination for collecting and visualizing cluster metrics.

The first step is to ensure that the metrics server is installed in the cluster, as it's responsible for gathering resource usage data like CPU and memory from nodes and pods.

Next, I deploy Prometheus in the cluster. Prometheus is responsible for scraping metrics from various Kubernetes components — such as the API server, kubelet, nodes, and even application-level metrics if they're exposed. I make sure Prometheus is configured to discover these targets automatically using Kubernetes service discovery.

After Prometheus is running, I set up Grafana, which is the visualization tool. Once Grafana is deployed, I connect it to Prometheus as the data source.

Then I import pre-built dashboards — such as ones available from the Grafana community — for Kubernetes monitoring. These dashboards give insights into things like node resource usage, pod health, deployments, cluster capacity, and overall system load.

Additionally, I make sure alerting is configured in Prometheus, so that if any threshold is crossed — like high CPU usage, pod restarts, or memory pressure — alerts are sent to tools like Slack, email, or PagerDuty.

In short, my setup process covers: metrics server installation, deploying Prometheus, connecting Grafana, importing dashboards, and setting up alerts to monitor the entire Kubernetes environment end to end.