Monitoring

How do you know if your cluster is healthy?

When running applications in Kubernetes observability is key. With K8ssandra and cass-operator, each Apache Cassandra pod is configured with the DataStax Metrics Collector for Apache Cassandra. This exposes Cassandra node-level metrics in the Prometheus format covering everything from operations per second and latency to compaction throughput and heap usage.

Grafana Overview

K8ssandra goes a step further providing a deployment of Prometheus and Grafana for the storage and visualization of these metrics including:

  • Prometheus ServiceMonitor resource with Metric Relabeling directed at the K8ssandra cluster’s Service
  • Grafana DataSource resource configured to reference the deployed Prometheus instance
  • Grafana Dashboard resources

Grafana Cluster

Metric Data Retention

Metric data collected in the cluster is retained within Prometheus for 24 hours.

Next

Check out Repairs with Cassandra Reaper



Last modified January 15, 2021: cass-operator upgrade (#188) (f213a4e)