-
-
Notifications
You must be signed in to change notification settings - Fork 759
Open
Labels
consumer groupKafka Consumer groupKafka Consumer groupenhancementNew feature or requestNew feature or requestneed investigation
Description
When using AKHQ to monitor a Kafka 4.1 cluster with consumer groups that use the new KIP-848 consumer rebalance protocol (group.protocol=consumer), the groups are incorrectly displayed with state DEAD in the UI, even though they are actively consuming and in a Stable state.
Environment
| Component | Version |
|---|---|
| AKHQ | 0.26 |
| Kafka | 4.1.0 |
| Deployment | Kubernetes (Strimzi Operator 0.48.0) |
| KRaft Mode | Yes |
Kafka Feature Flags:
group.version: 1metadata.version: 4.1-IV1
Broker Config:
group.coordinator.rebalance.protocols=classic,consumer
Expected Behavior
Consumer group should display state Stable (matching the actual state)
Actual Behavior
Consumer group displays state DEAD
Verification via CLI
The group is actually healthy when queried via CLI:
$ kafka-consumer-groups.sh --bootstrap-server localhost:9092 --describe --group <group-name> --state
GROUP COORDINATOR (ID) ASSIGNMENT-STRATEGY STATE #MEMBERS
<group-name> broker-0:9092 (0) uniform Stable 3
Additional CLI output shows KIP-848 specific fields (currentEpoch, targetEpoch):
$ kafka-consumer-groups.sh --describe --group <group-name> --members --verbose
GROUP CONSUMER-ID HOST CLIENT-ID #PARTITIONS CURRENT-EPOCH TARGET-EPOCH
<group-name> abc123... /10.x.x.x client-1 1 11 11 Additional Context
- Classic consumer groups (using the old protocol) display correctly in AKHQ
- Only groups using
group.protocol=consumerare affected - The issue is cosmetic but causes confusion when monitoring
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
consumer groupKafka Consumer groupKafka Consumer groupenhancementNew feature or requestNew feature or requestneed investigation
Projects
Status
Backlog