KafkaHow-ToBeginner · 4 min read

How to Monitor Kafka Cluster: Tools and Best Practices

To monitor a Kafka cluster, use its built-in JMX metrics exposed via Java Management Extensions, and collect them with tools like Prometheus. Visualize these metrics using dashboards such as Grafana or use Kafka-specific tools like Kafka Manager for cluster health and topic monitoring.

📐

Syntax

Kafka exposes metrics through JMX (Java Management Extensions). You enable JMX by setting environment variables when starting Kafka brokers. Metrics can be collected by monitoring tools using these settings:

JMX_PORT: Port where Kafka exposes metrics.
JMX_HOSTNAME: Hostname for JMX service.
KAFKA_JMX_OPTS: Java options to configure JMX behavior.

Example environment variable to enable JMX:

bash

export JMX_PORT=9999
export KAFKA_JMX_OPTS="-Dcom.sun.management.jmxremote -Dcom.sun.management.jmxremote.authenticate=false -Dcom.sun.management.jmxremote.ssl=false"

💻

Example

This example shows how to enable JMX on a Kafka broker and scrape metrics with Prometheus using the JMX Exporter. It demonstrates starting Kafka with JMX enabled and configuring Prometheus to collect metrics.

bash

# Step 1: Enable JMX on Kafka broker
export JMX_PORT=9999
export KAFKA_JMX_OPTS="-Dcom.sun.management.jmxremote -Dcom.sun.management.jmxremote.authenticate=false -Dcom.sun.management.jmxremote.ssl=false"

# Start Kafka broker (example command)
bin/kafka-server-start.sh config/server.properties

# Step 2: Configure Prometheus scrape config (prometheus.yml)
scrape_configs:
  - job_name: 'kafka'
    static_configs:
      - targets: ['localhost:9999']

# Step 3: Run Prometheus and Grafana to visualize metrics

Output

Kafka broker starts with JMX on port 9999 Prometheus scrapes metrics from localhost:9999 Grafana dashboard shows Kafka metrics like broker health, topic throughput, and consumer lag

⚠️

Common Pitfalls

Common mistakes when monitoring Kafka clusters include:

Not enabling JMX properly, so no metrics are exposed.
Using default JMX ports that conflict with other services.
Not securing JMX endpoints, exposing sensitive data.
Ignoring consumer lag metrics, which indicate processing delays.
Not setting up alerting on critical metrics like under-replicated partitions.

Always verify JMX ports and secure access in production.

bash

## Wrong: Starting Kafka without JMX enabled
bin/kafka-server-start.sh config/server.properties

## Right: Enable JMX before starting Kafka
export JMX_PORT=9999
bin/kafka-server-start.sh config/server.properties

📊

Quick Reference

Metric	Description	Why Monitor
Under-Replicated Partitions	Partitions not fully replicated	Indicates replication issues risking data loss
Consumer Lag	Delay of consumers behind producers	Shows if consumers are keeping up with data flow
Request Rate	Number of requests per second	Measures broker load and throughput
Offline Partitions	Partitions without leader	Shows cluster health and availability
Broker CPU and Memory	Resource usage of brokers	Detects performance bottlenecks

✅

Key Takeaways

Enable JMX on Kafka brokers to expose metrics for monitoring.

Use Prometheus and Grafana to collect and visualize Kafka metrics effectively.

Monitor critical metrics like under-replicated partitions and consumer lag to ensure cluster health.

Secure JMX endpoints to prevent unauthorized access.

Use Kafka Manager or similar tools for easier cluster and topic monitoring.