spuru9 opened a new pull request, #1189: URL: https://github.com/apache/flink-kubernetes-operator/pull/1189
## Summary - Follow-up to [FLINK-40443](https://issues.apache.org/jira/browse/FLINK-40443) / #1187, per maintainer consensus on that PR (separate PR, not blocking, does not touch the hardening fix itself). - Adds `KafkaPulsarPartitionMetricsITCase`, which verifies `ScalingMetricCollector#getJobTopology` derives the correct `numSourcePartitions` against real Kafka and Pulsar brokers (Testcontainers), covering: multi-topic summing, dotted and hyphenated topic names, source parallelism > 1, multiple source vertices in one job, and a non-Kafka/Pulsar source that must not report a partition count. - The Flink job under test runs via `MiniClusterJobDriver` in its own JVM rather than in-process: `flink-autoscaler` vendors a compatibility copy of `org.apache.flink.runtime.rest.messages.job.JobDetailsInfo` under `src/main/java` that shadows the real Flink class on this module's own classpath, which breaks a same-process JobManager's own REST handlers. - Wires the new test into CI (`ci.yml`) with the same per-module `verify -Dit.skip=false` step already used for `flink-kubernetes-operator`, `flink-kubernetes-webhook`, and `flink-autoscaler-plugin-jdbc`. No JIRA ticket filed yet — opening as draft for early feedback; happy to file one before this is ready for review. ## Test plan - [x] `mvn -pl flink-autoscaler verify -Dit.skip=false` passes locally (290 unit tests + the new IT) - [x] `mvn -pl flink-autoscaler checkstyle:check spotless:check` clean - [ ] CI run on this PR 🤖 Generated with [Claude Code](https://claude.com/claude-code) -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
