Kafka Consumer: Reading Messages from a Topic
A busy Kafka topic holds far more than you need for one investigation. The Console reads a bounded window directly from Kafka, then lets you narrow the result or switch to a streaming view.
What is a Kafka consumer?
A Kafka consumer reads records from topic partitions and tracks its position
with an offset for each partition. Apache Kafka includes the
kafka-console-consumer command for reading records in a terminal.
Kafma reads directly from the brokers for interactive inspection. Opening or refreshing a Console tab does not join a consumer group, trigger a rebalance, or commit offsets. To inspect application consumers and their committed positions, see Consumer groups.
Consume from the beginning or a timestamp
Choose From based on the part of the topic you need:
- Newest loads a one-time snapshot of the latest records.
- Beginning starts at the earliest records still retained by Kafka. Records already removed by retention are not available.
- Offset starts each partition at the specified offset.
- Timestamp starts each partition at its first available record at or after the selected local date and time.
Limit caps the records loaded by a snapshot. It keeps a large topic bounded without changing anything on the broker.
Search and inspect Kafka messages
Search filters the records already loaded into the Console; it does not scan the entire topic. Enter text to search message keys and values. Use the partition filter when offsets or ordering need to be read in one partition.
Expand a row to see its full key and value, headers, partition, offset, timestamp, size, and compression. Avro and Protobuf payloads decode automatically when a Schema Registry is configured.
Bulk Forward and Export use the current filtered set. Narrow the view before reusing messages.
Live and Watch Group modes
Live starts at the topic's current end and shows records produced from that point forward. Its buffer holds the latest 500 records; older records dropped from the view remain on the broker.
Watch Group shows a bounded window around a selected consumer group's committed positions. It marks records as consumed (✓✓), pending (◷), or unassigned (·), and shows the group's member status and lag. Kafma reads this information without joining the group.
The divider separates the group's remaining backlog above from its most recently consumed record below. The lag count and trend show whether the group is catching up or falling further behind, while the messages at the front of the lag help you investigate stalled consumption.
