y1chi commented on code in PR #26085:
URL: https://github.com/apache/beam/pull/26085#discussion_r1178393788
##########
runners/google-cloud-dataflow-java/worker/src/main/java/org/apache/beam/runners/dataflow/worker/windmill/GrpcWindmillServer.java:
##########
@@ -987,6 +998,84 @@ public void append(StreamingGetWorkResponseChunk chunk) {
this.data = data.concat(chunk.getSerializedWorkItem());
this.bufferedSize += chunk.getSerializedWorkItem().size();
+ for (GetWorkStreamTimingInfo info :
chunk.getPerWorkItemTimingInfosList()) {
+ getWorkStreamTimings.compute(
+ info.getEvent(),
+ (event, recordedTime) -> {
+ Instant newTimingForEvent =
Instant.ofEpochMilli(info.getTimestampUsec() / 1000);
Review Comment:
Added comments. I agree that we probably want to identifying the max elapsed
time with each stage, which at least tells us where exactly the slowness is.
I've changed implementation to take that.
--
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]
