Kalvin2077 opened a new pull request, #3858: URL: https://github.com/apache/celeborn/pull/3858
### What changes were proposed in this pull request? Sub PR of [CIP-13](https://cwiki.apache.org/confluence/spaces/CELEBORN/pages/325421875/CIP-13+Support+automatic+scaling) This PR adds the worker metrics foundation for automatic scaling: - Adds a typed `WorkerStats` model and protobuf messages. - Adds a `ScaleMetricCollector` extension point. - Adds a default collector for memory, system load, and local disk metrics. - Reports the latest worker stats in worker heartbeats. - Adds a master-side temporary `WorkerStatsStore`. - Adds configuration keys and documentation. - Keeps metrics disabled by default. ### Why are the changes needed? Automatic scaling requires reliable worker load input. The current worker heartbeat contains lifecycle state and disk information, but it does not carry resource usage metrics. This PR provides that input without implementing the scaling policy or platform operations. ### Does this PR resolve a correctness bug? - [ ] Yes ### Does this PR introduce _any_ user-facing change? - [x] Yes Adds new configuration keys for enabling and controlling worker scale metrics. The feature is disabled by default. ### How was this patch tested? - `WorkerStatsSuite` verifies protobuf round-trip and metric status semantics. - `DefaultScaleMetricCollectorSuite` verifies memory, system load, and disk metric collection. - `WorkerStatsStoreSuite` verifies latest-snapshot retention and cleanup. - Ran `spotless:check`. - Ran the module build with JDK 17. -- 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]
