Rangsh commented on PR #12173: URL: https://github.com/apache/seatunnel/pull/12173#issuecomment-5628982343
Thanks @SEZ9 — aligning against current head `c553a8c04`: 1. **Durability sample (F1 / F3)** — Agreed on the mid-trial gap. Documented in `IMapJobGrowthBenchmarkWorkload` / `IMapJobStorageBenchmark` Javadoc (`c553a8c04`): iterations between the first empty-pressure sample and trial tear-down are resident-only, so a WAL append that starts failing in that window surfaces at tear-down (resident size checks still run every iteration). Full-batch `loadAll` widening tracked as follow-up: https://github.com/apache/seatunnel/issues/12266 2. **`storeFinishedPipelineMetrics` atomicity (F2)** — Confirmed in `ca82f95`: `finishedJobMetricsImap.lock(jobId)` / `unlock(jobId)` around `get → merge → put` with `try`/`finally`. 3. **Docs (F4)** — Per @nzw921rx, the growth methodology notes were removed again from `docs/en|zh/engines/zeta/benchmark.md` in `9a6077c58` (benchmark docs already refactored; per-scenario detail stays in Javadoc only). 4. **F5 Javadoc** — Done in `ca82f95`: contract is lock + merge + single TTL `put`; no historical `computeIfAbsent` / “in memory” rationale. 5. **F6 tests** — Done in `ca82f95`: `JobHistoryServiceFinishedMetricsTest` asserts observable IMap `put` / merge / null rejection / lock-unlock; no `never().computeIfAbsent`. Please take another look when you have a chance. @nzw921rx — docs removal is in `9a6077c58`; no further doc changes in this push. Thanks for waiting. -- 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]
