Cassandra OpenTelemetry Assets


Version	0.1.0 (View all)
Subscription level What's this?	Basic
Developed by What's this?	Elastic
Minimum Kibana version(s)	9.2.1

The Cassandra OpenTelemetry Assets integration v0.1.0 is in beta

To use beta integrations, go to the Integrations page in Kibana, scroll down, and toggle on the Display beta integrations option.

Apache Cassandra is a distributed, wide-column NoSQL database designed for high availability and linear scalability across commodity hardware. The Cassandra OpenTelemetry assets provide dashboards, alert rules, and SLO templates for metrics collected using the OpenTelemetry JMX receiver with target_system: cassandra, covering latency, traffic, errors, compaction, and storage across your cluster.

Compatibility

The Cassandra OpenTelemetry assets have been tested with:

OpenTelemetry JMX Scraper from opentelemetry-java-contrib v1.54.0 with EDOT Collector v9.2.1
OpenTelemetry JMX Scraper from opentelemetry-java-contrib v1.54.0 with OpenTelemetry Collector Contrib v0.146.0

Cassandra tested against:

Apache Cassandra 4.1

Requirements

You need Elasticsearch for storing and searching your data and Kibana for visualizing and managing it. You can use our hosted Elasticsearch Service on Elastic Cloud, which is recommended, or self-manage the Elastic Stack on your own hardware.

Setup

Prerequisites

Cassandra exposes metrics using JMX on port 7199 by default. Ensure JMX remote access is enabled and reachable from where you run the collector. If you use the standalone JMX Scraper, run it with OTEL_JMX_TARGET_SYSTEM=cassandra and configure it to export using OTLP to your collector endpoint.

Configuration

Install and configure the upstream OpenTelemetry Collector or Elastic Distribution of OpenTelemetry (EDOT) Collector to receive metrics from the JMX receiver or JMX Scraper and export them to Elasticsearch. Replace the following placeholders in the configuration:

Placeholder	Description	Example
`<ES_ENDPOINT>`	Elasticsearch endpoint	`https://elasticsearch.example.com:9200`
`<ES_API_KEY>`	Elasticsearch API key for authentication	`${env:ES_API_KEY}`
`<CASSANDRA_JMX_ENDPOINT>`	Cassandra JMX endpoint (host:port or full JMX URL)	`cassandra-host:7199`
`<JMX_JAR_PATH>`	Path to the JMX Metric Gatherer or JMX Scraper JAR	`/opt/opentelemetry-java-contrib-jmx-metrics.jar`

		receivers:
  jmx/cassandra:
    jar_path: <JMX_JAR_PATH>
    endpoint: <CASSANDRA_JMX_ENDPOINT>
    target_system: cassandra
    collection_interval: 10s

processors:
  resource/dataset:
    attributes:
      - key: data_stream.dataset
        value: cassandra
        action: upsert
  batch:
    timeout: 10s
    send_batch_size: 1024

exporters:
  elasticsearch/otel:
    endpoint: <ES_ENDPOINT>
    api_key: <ES_API_KEY>
    mapping:
      mode: otel

service:
  pipelines:
    metrics:
      receivers: [jmx/cassandra]
      processors: [resource/dataset, batch]
      exporters: [elasticsearch/otel]
		
	

If you use the standalone JMX Scraper instead of the JMX receiver, replace the jmx/cassandra receiver with an otlp receiver and point the JMX Scraper's OTLP exporter to your collector. Ensure the scraper runs with OTEL_JMX_TARGET_SYSTEM=cassandra.

Note

For multi-node clusters, run one JMX receiver or scraper instance per Cassandra node, or use a single collector with multiple JMX receiver configs, each pointing to a different node.

Dashboard	Description
[Cassandra OTel] Overview	Overview of Apache Cassandra cluster health from OTel JMX metrics: client request latency, throughput, errors, compaction, and storage.

Alert rules

Alert	Trigger	Severity
[Cassandra OTel] High read latency (p99)	p99 read latency exceeds 100 ms on any node	High
[Cassandra OTel] High write latency (p99)	p99 write latency exceeds 50 ms on any node	High
[Cassandra OTel] High range slice latency (p99)	p99 range slice latency exceeds 500 ms on any node	High
[Cassandra OTel] Request errors (Timeout, Unavailable, Failure)	Any request errors detected (Timeout, Unavailable, or Failure)	Critical
[Cassandra OTel] High error rate by node	Error rate exceeds 5% on any node	Critical
[Cassandra OTel] Compaction falling behind (pending tasks)	Pending compaction tasks greater than zero on any node	Warning
[Cassandra OTel] Hints in progress (replicas unreachable)	Hints in progress greater than zero on any node	Warning

SLO templates

Note

SLO templates require Elastic Stack version 9.4.0 or later.

SLO	Target	Window	Description
[Cassandra OTel] Read latency p99 99.5% rolling 30 days	99.5%	30-day rolling	p99 read latency below 50 ms for 99.5% of 1-minute intervals.
[Cassandra OTel] Request success rate 99.5% rolling 30 days	99.5%	30-day rolling	Success rate (total requests minus errors) at least 99.5% for 99.5% of 1-minute intervals.

Screenshots

This integration includes one or more Kibana dashboards that visualizes the data collected by the integration. The screenshots below illustrate how the ingested data is displayed.