pvary commented on code in PR #23737:
URL: https://github.com/apache/flink/pull/23737#discussion_r1398978869
##########
flink-tests/src/test/java/org/apache/flink/test/streaming/runtime/SinkV2MetricsITCase.java:
##########
@@ -217,43 +215,41 @@ private void assertSinkMetrics(
assertThat(subtaskWithMetrics, equalTo(numSplits));
}
- private void assertSinkCommitterMetrics(
- JobID jobId, int parallelism, Map<String, Long> expected) {
+ private void assertSinkCommitterMetrics(JobID jobId, Map<String, Long>
expected) {
List<OperatorMetricGroup> groups =
reporter.findOperatorMetricGroups(
jobId, TEST_SINK_NAME + ": " + DEFAULT_COMMITTER_NAME);
- assertThat(groups, hasSize(parallelism));
Review Comment:
I originally thought, that when starting a job the TMs/Operators are
created, and we start serving records only when the full infrastructure is
ready.
Turns out, that the TMs start serving records whenever they are ready, and
there is no synchronisation point between them. This caused the flakiness of
the test.
I have found that 1 out of 10.000 cases not even the metrics group is
created when we arrive at this point.
Since the goal is to check that the values of the metrics are correct, I
decided to remove this assertion.
--
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]
