Apache Cassandra integration

The Apache Cassandra integration provides cluster monitoring for Cassandra, covering latency, compaction, JVM performance, and per-keyspace metrics. Cassandra is distributed, so monitoring spans the whole cluster, from read and write latency to compaction backlog and the JVM health of each node.


What it’s for	Monitoring cluster health, read/write latency, compaction, and JVM performance
Under the hood	The JMX exporter runs on each node. Grafana Alloy scrapes the metrics it exposes

Metrics collected

Latency: Cluster-wide and per-node read and write latency
Compaction: Pending and blocked compaction tasks
JVM: Garbage collection and memory
Cluster: Node status (up/down) and endpoint health
Keyspaces: Repair jobs, pending compactions, and disk usage

What to know

This integration offers:

Prebuilt dashboards: overview, nodes, keyspaces, logs
Prebuilt alerts: read/write latency, blocked compactions, CPU/memory

Prerequisites:

Requires the JMX exporter on each Cassandra node

Documentation

View the Apache Cassandra integration documentation.

Cassandra is a distributed database, so its monitoring spans the whole cluster: read and write latency, compaction backlog, and the health of each node’s JVM.

Setup is a little more involved than other integrations because Cassandra exposes its metrics through JMX. You run the JMX exporter on each node, and Alloy scrapes what it exposes. The payoff is visibility into compaction, garbage collection, and repair health across the cluster.

Apache Cassandra integration