[
https://issues.apache.org/jira/browse/CASSANDRA-1380?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12897452#action_12897452
]
Kelley Reynolds commented on CASSANDRA-1380:
--------------------------------------------
If your entire cluster is dead then yes, you won't be able to access the data
inside of it but that's only one extreme case under which it's not useful.
Trended statistics are attractive for much more than just monitoring. You could
see impending I/O bottlenecks, RF>1 clusters would let you examine behavior
just before a crash, facilitate debugging and recommendations, and myriad other
things that other people can think of. Many other systems have stats collectors
built-in that are used to augment existing monitoring systems and
troubleshoot/optimize application behavior, I can't think of a reason Cassandra
should be any different.
> Create a statistics keyspace
> ----------------------------
>
> Key: CASSANDRA-1380
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1380
> Project: Cassandra
> Issue Type: New Feature
> Reporter: Kelley Reynolds
> Priority: Minor
>
> Most significant installations of cassandra have monitoring systems that
> monitor and trend several key tpstats among other things. Instead of
> replicating that for each and every installation, it would be useful to have
> that provided by the core system in a default keyspace. Comprehensive
> statistics would allow for improved debugging tools, performance
> recommendations, helpful alerts, introspection, that sort of thing.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
