AHeise commented on a change in pull request #10931: [FLINK-15603][metrics] Add
checkpointStartDelay metric
URL: https://github.com/apache/flink/pull/10931#discussion_r370092256
##########
File path:
flink-streaming-java/src/main/java/org/apache/flink/streaming/runtime/io/CheckpointBarrierHandler.java
##########
@@ -95,4 +101,10 @@ protected void notifyAbortOnCancellationBarrier(long
checkpointId) throws Except
protected void notifyAbort(long checkpointId, CheckpointException
cause) throws Exception {
toNotifyOnCheckpoint.abortCheckpointOnBarrier(checkpointId,
cause);
}
+
+ protected void markCheckpointStart(long checkpointCreationTimestamp) {
+ latestCheckpointStartDelayNanos = 1000_000 * Math.max(
Review comment:
Yes, I'd convert to ms. I don't see why any user profits from ns.
If it was some per-record metric, I'd agree with nano (like in Kinesis?).
It's technically much too complicated to actually produce nano-precision
given the expected usability gain.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
