[
https://issues.apache.org/jira/browse/FLINK-9665?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16524939#comment-16524939
]
ASF GitHub Bot commented on FLINK-9665:
---------------------------------------
Github user zentol commented on a diff in the pull request:
https://github.com/apache/flink/pull/6211#discussion_r198460116
--- Diff:
flink-metrics/flink-metrics-prometheus/src/test/java/org/apache/flink/metrics/prometheus/PrometheusReporterTest.java
---
@@ -159,6 +162,27 @@ public void histogramIsReportedAsPrometheusSummary()
throws UnirestException {
}
}
+ @Test
+ public void metricIsRemovedWhenCollectorIsNotUnregisteredYet() throws
UnirestException {
+ TaskManagerMetricGroup tmMetricGroup = new
TaskManagerMetricGroup(registry, HOST_NAME, TASK_MANAGER);
+
+ String metricName = "numRecordsOut";
+
+ Counter metric1 = new SimpleCounter();
+ FrontMetricGroup<TaskManagerJobMetricGroup> metricGroup1 = new
FrontMetricGroup<>(1, new TaskManagerJobMetricGroup(registry, tmMetricGroup,
JobID.generate(), "job_1"));
+ reporter.notifyOfAddedMetric(metric1, metricName, metricGroup1);
+
+ Counter metric2 = new SimpleCounter();
+ FrontMetricGroup<TaskManagerJobMetricGroup> metricGroup2 = new
FrontMetricGroup<>(2, new TaskManagerJobMetricGroup(registry, tmMetricGroup,
JobID.generate(), "job_2"));
--- End diff --
nit: the first argument to `FrontMetricGroup` should be identical since
you're passing it to the same reporter. Actually it shouldn't be necessary to
create a `FrontMetricGroup` at all, as it is only need to introduce reporter
specific configuration properties.
> PrometheusReporter does not properly unregister metrics
> -------------------------------------------------------
>
> Key: FLINK-9665
> URL: https://issues.apache.org/jira/browse/FLINK-9665
> Project: Flink
> Issue Type: Bug
> Components: Metrics
> Affects Versions: 1.5.0, 1.4.2, 1.6.0
> Reporter: Chesnay Schepler
> Assignee: Jelmer Kuperus
> Priority: Major
> Labels: pull-request-available
>
> The {{PrometheusReporter}} groups metrics with the same logical scope in a
> single {{Collector}} which are periodically polled by Prometheus.
> New metrics are added to an existing collector, and a reference count is
> maintained so we can eventually cleanup the {{Collector}} itself.
> For removed metrics we decrease the reference count, do not however remove
> the metrics that were added. As a result the collector will continue to
> expose metrics, as long as at least 1 metric exists with the same logical
> scope.
> If the collector is a {{io.prometheus.client.Gauge}} we can use the
> {{#remove()}} method. For histograms we will have to modify our
> {{HistogramSummaryProxy}} class to allow removing individual histograms.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
