[
https://issues.apache.org/jira/browse/SOLR-10262?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15985072#comment-15985072
]
Andrzej Bialecki commented on SOLR-10262:
------------------------------------------
[~otis] - actually, I hesitated to add either HdrHistogram or the one from
rolling-metrics. The first one doesn't implement exponential decay or at least
some form of incremental reset (eg. time-window based), and reset-on-snapshot
will never work correctly with multiple reporters. The rolling-metrics package
implements exponential decay and much more, but it does it by using multiple
partial histograms for each Histogram instance, which concerns me from the
point of view of overall memory impact (we create hundreds of metrics per
core), background thread mgmt, etc.
So, from my point of view the patch as it is now may be sufficient and maybe we
should stop here - it allows you to customize what implementation and
parameters of each metric is used, and it allows you to provide your own
implementation and evaluate its impact. Please review the patch and let me know
what you think.
> Collect request latency metrics for histograms
> ----------------------------------------------
>
> Key: SOLR-10262
> URL: https://issues.apache.org/jira/browse/SOLR-10262
> Project: Solr
> Issue Type: New Feature
> Security Level: Public(Default Security Level. Issues are Public)
> Components: metrics
> Reporter: Otis Gospodnetic
> Assignee: Andrzej Bialecki
> Fix For: master (7.0)
>
> Attachments: SOLR-10262.patch
>
>
> Since [~ab] is on a role with metrics...
> There is no way to accurately compute request latency percentiles from
> metrics exposed by Solr today. We should consider making that possible. c.f.
> https://github.com/HdrHistogram/HdrHistogram
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
