I'm managing a small Kubernetes cluster with about 4 nodes. Currently, I'm using Grafana Cloud along with Alloy, deployed as a DaemonSet for metrics and log collection. However, I'm finding it a bit clunky for my needs. I'm thinking about switching to the kube-prometheus-stack but I'm not entirely sure. What tools do you guys use for monitoring and alerting, and what are the advantages of those tools?
3 Answers
Prometheus is definitely a solid choice for monitoring Kubernetes. It's widely recommended and integrates well with other tools like Grafana. If you're already using Grafana, sticking with Prometheus will give you access to a ton of preconfigured dashboards that can help you get started quickly.
We use direct integration of checkmk for monitoring multiple Kubernetes clusters, and it works really well. You can add Prometheus metrics for a deeper view using PromQL. With checkmk, you'll get comprehensive monitoring of resources, including nodes and deployments, and if you install the checkmk agent on the nodes, it also provides additional server metrics.
You really only need Prometheus combined with Alertmanager and Grafana, plus Karma for alert dashboards. This combo should cover all your bases for monitoring and alerting effectively.
Is kube-prometheus-stack a good option then? Can I set up alerts with it?