featzhang commented on PR #27774:
URL: https://github.com/apache/flink/pull/27774#issuecomment-4365407494

   Self-review pass (amended into the second commit, HEAD `d2449326c60`):
   
   **API polish**
   - Removed the dead backwards-compatible 3-arg `TopNMetricsResponseBody` 
constructor; every call-site was already on the 5-arg form.
   - All five top-level list fields are now null-guarded consistently 
(previously only the two new fields were).
   - `CpuConsumerInfo.subtaskId` is now `Integer` and reported as `null` for 
TaskManager-scoped CPU readings, matching the `null`-means-unavailable 
convention used by `SourceLagInfo`. Sentinel `-1` is gone.
   - `SourceLagInfo` javadoc and the TS interface are aligned on `null` 
(previously claimed `-1`).
   - `TopNMetricsHeaders.getDescription()` now enumerates the five actual 
dimensions instead of the stale 3-dimension wording that was leaking into the 
OpenAPI doc.
   
   **Handler simplification**
   - `findOperatorMetric`: dropped the never-taken bare-key fallback; only the 
operator-scoped `".<metricName>"` suffix is matched now, with an updated 
comment tying the behaviour to FLIP-33.
   - Removed the `getFriendlyTaskManagerName` helper that read `taskmanager.id` 
/ `taskmanager.host` — no such metrics are exported anywhere in the tree, so 
the helper always fell through to `tmId`.
   - Source-lag aggregation uses `Double maxFetchLag / maxEmitLag = null` 
instead of `Double.NEGATIVE_INFINITY` sentinels.
   - Repeated `?topN=…` query parameters now follow the "last value wins" HTTP 
convention.
   
   **REST snapshot**
   - `rest_api_v1.snapshot` regenerated to actually include `topBusyOperators` 
and `topLaggingSources` (the snapshot had been missed when those fields were 
first added).
   
   Verified locally: `spotless:check`, `checkstyle:check`, `TopNMetrics*` tests 
(20/20 green), and `RuntimeRestAPIStabilityTest` pass. No behavioural change 
for callers that already consumed the 5-dimension response.
   


-- 
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]

Reply via email to