JeremyXin commented on PR #12169: URL: https://github.com/apache/seatunnel/pull/12169#issuecomment-6386597946
@DanielLeens @nzw921rx Thank you both for the detailed reviews and for staying with this PR through the updates. I have addressed the two architecture questions on the current head. The autoscaler signal store is now owned by CoordinatorService; ResourceManager remains responsible for worker registration and slot allocation and reports capacity shortages through a ResourceShortageListener. I also moved AutoscalerConfig into EngineConfig, so the master and workers obtain it from the engine configuration instead of passing a separate config through the startup constructors. On the E2E request: Slice 1 only evaluates signals and publishes advisory recommendations. It does not create or remove workers, and there is no external consumer of the recommendation yet. The policy, state transitions, publication, and coordinator lifecycle have unit test coverage. Would you be comfortable deferring an E2E that verifies an actual worker change to the slice that adds that behavior? If you want integration coverage of the advisory path in this PR, I can add that separately. -- 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]
