Rafał Harabień created SOLR-17276:
-------------------------------------
Summary: Use fixed rate instead of fixed delay in
prometheus-exporter
Key: SOLR-17276
URL: https://issues.apache.org/jira/browse/SOLR-17276
Project: Solr
Issue Type: Improvement
Security Level: Public (Default Security Level. Issues are Public)
Components: contrib - prometheus-exporter
Reporter: Rafał Harabień
Attachments: image-2024-05-06-18-10-30-739.png
Prometheus-exporter is supposed to read metrics from Solr servers every 60
seconds (scrape interval can be changed using --scrape-interval argument). But
the truth is it does it every 60+X seconds where X is the time needed to read
metrics from all Solr servers. In my case X is 1-2 s. If Prometheus scrapes the
exporter every 60 seconds it can lead to duplicated samples (e.g. metrics will
stay the same for 2 minutes).
!image-2024-05-06-18-10-30-739.png!
It's result of using
[scheduler.scheduleWithFixedDelay|https://github.com/apache/solr/blob/2bb2ada0a372f4d101b78df8d43e0fc44c8edbf3/solr/prometheus-exporter/src/java/org/apache/solr/prometheus/collector/SchedulerMetricsCollector.java#L77]
instead of
scheduler.scheduleAtFixedRate.
Note: function scheduled with scheduleAtFixedRate can still be started late if
previous execution has not finished. There is no risk of overlapping executions.
I am going to prepare a PR.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
