Browse and Inspect Kafka Topics
What is a Kafka topic?
A Kafka topic is a named stream of records. Producers append records to the topic, and consumers read them without removing them.
Kafka divides a topic into partitions. Records are ordered within each partition and identified by offsets. Partitions distribute storage and processing across brokers, while additional replicas can keep copies available when a broker fails.
Multiple consumer groups can read the same topic at their own pace. Kafka keeps records according to the topic's retention and cleanup policy, so one consumer group does not control what another group can read.
Browse and filter Kafka topics
The Topics page lists every topic on the connected cluster. Filter by name or cleanup policy, and choose whether to include internal topics. Each row shows the topic's health, partition and replica counts, retention, and message count.
Inspect Kafka topic details
Select a topic to open its detail drawer.
Health, retention, and message count
Status follows the least healthy partition. Healthy means every partition has a leader and all replicas are in sync. Under Replicated means every partition still has a leader, but at least one is missing an in-sync replica. Offline means at least one partition has no elected leader; that partition cannot serve reads or writes.
The drawer shows ISR as the lowest in-sync replica count across the topic's partitions, compared with its replication factor.
Messages is the sum of the high-minus-low offset span for every partition. It
represents retained records on delete-only topics. Compaction can leave gaps, so
Kafma prefixes the count with ~ on compacted and mixed-policy topics. The value
below it is the topic's message size limit (max.message.bytes).
Partitions shows the partition count and how many different brokers currently lead them.
Retention summarises the cleanup policy and the retention window.
Partition offsets and leaders
Each partition lists its low and high offsets. The low offset is the earliest
record still available; the high offset is the next offset Kafka will assign.
Their difference is the displayed message span. L: 1 identifies the broker
currently leading the partition.
Compare the spans to spot uneven distribution. A much larger span can indicate skewed keys or an uneven partitioning strategy. The low offset is not always zero: it moves when retention removes old segments or after records are deleted.
Consumer groups
The Consumers section lists groups with active assignments or committed offsets for this topic. Each row shows the group state, topic-specific lag, and member count.
Select a consumer group to open its full details in the same drawer. Use the back button to return to the topic without leaving the Topics page, keeping topic and consumer diagnostics in one workflow.
Pin a topic to the Watchlist
Pinning a topic adds it to the Watchlist on Home, alongside pinned schema subjects and consumer groups. The Watchlist spans clusters, so a production topic and its local counterpart can sit side by side.
Delete a Kafka topic
Deletion is permanent and asynchronous: Kafka marks the topic and removes its data
in the background, so the topic may linger briefly in listings. It also requires
delete.topic.enable on the brokers, which is on by default in current versions.
Kafka removes committed offsets for the deleted topic as deletion propagates. Schema Registry subjects are separate and remain until you delete them.
Deletion is blocked entirely on connections marked read-only.