Inspect Kafka Brokers and Cluster Configuration
Kafma brings each broker and its configuration into one read-only view. Use it to confirm which nodes are reachable, find the controller, and spot settings that differ between brokers.
What is a Kafka broker?
A Kafka broker is a server that stores topic partitions and handles requests from producers and consumers. A Kafka cluster contains one or more brokers and has an active controller for metadata operations.
Inspect Kafka brokers
The header shows the cluster ID, metadata mode, and connection security. Each broker card shows its node ID, endpoint, controller role, storage, and number of partition leaders.
Storage is the current size of the partition logs on that broker, not its disk capacity or free space.
On a multi-broker cluster, the percentages beside Storage and Leaders compare that broker with the cluster average. Large differences can point to an uneven distribution that needs investigation.
The reachability indicator tests the broker's TCP endpoint. It confirms whether Kafma can connect to that address; it is not a complete broker health check.

Read Kafka broker configuration
Cluster Config shows settings with the same value across the displayed brokers. Choose a broker from the menu, or click its card, to inspect that broker's values.
Use Overrides only to hide Kafka defaults and focus on explicitly configured values. Search by configuration name, and hover over truncated names or values to see them in full. Sensitive values remain redacted. Kafma does not edit broker configuration from this page.
Hover a Source badge to see whether the value comes from Kafka defaults, static broker configuration, a cluster-wide default, a per-broker override, or mixed sources.
Find divergent broker settings
Divergent Config appears when a setting has different values across brokers or is unavailable on one of them. It places each broker's value in the same row so you can compare them directly.
Some differences, such as broker.id, node.id, broker.rack, or listeners,
are expected. Use this view to find unexpected differences in settings that
should be consistent across the cluster.
