Runbook #9 – Getting Started with Cassandra Troubleshooting and Starting the Service

This article provides insights on troubleshooting a Cassandra cluster when issues occur. The first step is to monitor the cluster’s health using various monitoring solutions such as Dynatrace, New Relic, and Prometheus\Grafana. Monitoring dashboards are used to identify anomalies in metrics like CPU utilization, read\write throughput, latency, read\write timeouts, and network utilization. Log aggregation tools like Logstash, Fluentd, and Graylog are used to search for problematic logs across all nodes of the cluster. Nodetool commands are used to get detailed information about what is going on with the cluster. Lastly, OS tools are used to identify what is happening at the OS/Hardware level. The article emphasizes the importance of looking at the graphs and logs from all nodes collectively to understand if the issue is isolated or cluster-wide.

Excerpted from the text – “A Cassandra cluster is an excellent tool for storing, accessing and managing high-volume data that is highly available and reliable.”