Explore Kafka metrics
Understanding how to interpret the key Kafka performance metrics displayed in your dashboards is essential for maintaining a healthy cluster. Important metrics include message throughput, byte rates, consumer lag, under-replicated partitions, and request latencies.
Normal operating ranges vary by workload, but certain metrics have clear indicators of problems. Under-replicated partitions should typically be 0 in a healthy cluster, the active controller count should always be 1, and consumer lag should remain low relative to your message ingestion rate.
Navigate to Dashboards from the main menu.
Find and open the Kafka dashboards folder, then open the Kafka Overview dashboard.
Review the Messages In Rate panel to see the volume of messages being produced to your topics.
Check the Bytes In/Out Rate panels to monitor data throughput across your cluster.
Examine the Under Replicated Partitions panel — this should typically be 0 in a healthy cluster.
Review the Request Latency panels to identify slow produce or fetch requests.
Check the Active Controller Count panel — this should always be 1 in a healthy cluster.
Navigate to the Kafka Broker Metrics dashboard to view individual broker performance, including CPU usage, network throughput, and request queue size.
Explore the Kafka Topic Metrics dashboard to analyze per-topic message rates and retention.
Use the time range selector to analyze historical trends and identify patterns.
You can now interpret the Kafka metrics displayed across your dashboards and identify which panels indicate healthy operation versus potential issues.
In the next milestone, you celebrate completing this learning journey!