cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Ryan King (JIRA)" <j...@apache.org>
Subject [jira] Updated: (CASSANDRA-1024) Latency Histograms
Date Tue, 27 Apr 2010 22:17:32 GMT

     [ https://issues.apache.org/jira/browse/CASSANDRA-1024?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Ryan King updated CASSANDRA-1024:
---------------------------------

    Attachment: 1024.patch

This is w WIP patch. Current it does:

* keeps estimated histograms on a geometric scale from 1 microsecond to around 30 seconds
(each bucket is 1.2 larger than the previous)
* exposes read and write latency histograms for each CF (not sure if this is the best way)

TODO:

* recent histograms
* deal with some concurrency issues (should probably use AtomicLongArray's)
* percentiles (export useful percentiles (25, 50, 75, 90, 99, 99.9, etc))

> Latency Histograms
> ------------------
>
>                 Key: CASSANDRA-1024
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-1024
>             Project: Cassandra
>          Issue Type: Improvement
>            Reporter: Ryan King
>            Assignee: Ryan King
>         Attachments: 1024.patch
>
>
> In addition to total and recent latency, I would like to keep track of histograms. We
have a way for doing estimated histograms (whose error domain gets worse the higher the latency
and whose storage requirements are constant). It works pretty well for us:
>   http://github.com/robey/ostrich/blob/master/src/main/scala/com/twitter/ostrich/Histogram.scala
> I'm planning on doing the work, but wanted to know if it would be useful to others.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message