Issue submitter TODO list
Describe the bug (actual behavior)
Summary
Upgrading from v1.4.2 → v1.5.0 causes Confluent Cloud clusters to become completely inaccessible. The root cause is a combination of two things: StatisticsService always calls describeLogDirs() (unsupported on Confluent Cloud), and v1.5.0 introduced AdminClient invalidation on ClusterAuthorizationException, which was not present in v1.4.2.
Root cause
ScrapedClusterState.scrape() unconditionally calls ac.describeLogDirs(). Confluent Cloud is a managed service and does not support this admin API, returning ClusterAuthorizationException.
In v1.5.0, StatisticsService.getStatistics() added:
.doOnError(e -> adminClientService.invalidate(cluster, e))
And AdminClientServiceImpl.invalidate() kills the AdminClient for any org.apache.kafka.common.errors.* exception — including ClusterAuthorizationException. This makes the entire cluster unreachable.
In v1.4.2 the invalidation did not exist, so the error was logged and the cluster continued to function.
Expected behavior
No response
Your installation details
- Docker compose
- Kafka-UI version 1.5.0
- Confluent cloud environment
Steps to reproduce
- Configure kafbat-ui against a Confluent Cloud cluster with
SASL_SSL / PLAIN
- Run
v1.4.2 → works fine
- Upgrade to
v1.5.0 → cluster immediately shows as offline
Screenshots
No response
Logs
ERROR io.kafbat.ui.service.StatisticsService : Failed to collect cluster <name> info
org.apache.kafka.common.errors.ClusterAuthorizationException: Cluster authorization failed.
WARN io.kafbat.ui.service.AdminClientServiceImpl : AdminClient for the cluster <name> is invalidated due to Cluster authorization failed.
Additional context
No response
Issue submitter TODO list
main-labeled docker image and the issue still persists thereDescribe the bug (actual behavior)
Summary
Upgrading from
v1.4.2→v1.5.0causes Confluent Cloud clusters to become completely inaccessible. The root cause is a combination of two things:StatisticsServicealways callsdescribeLogDirs()(unsupported on Confluent Cloud), and v1.5.0 introduced AdminClient invalidation onClusterAuthorizationException, which was not present in v1.4.2.Root cause
ScrapedClusterState.scrape()unconditionally callsac.describeLogDirs(). Confluent Cloud is a managed service and does not support this admin API, returningClusterAuthorizationException.In v1.5.0,
StatisticsService.getStatistics()added:And
AdminClientServiceImpl.invalidate()kills the AdminClient for anyorg.apache.kafka.common.errors.*exception — includingClusterAuthorizationException. This makes the entire cluster unreachable.In v1.4.2 the invalidation did not exist, so the error was logged and the cluster continued to function.
Expected behavior
No response
Your installation details
Steps to reproduce
SASL_SSL/PLAINv1.4.2→ works finev1.5.0→ cluster immediately shows as offlineScreenshots
No response
Logs
Additional context
No response