[
https://issues.apache.org/jira/browse/CASSANDRA-14537?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16520227#comment-16520227
]
Benedict commented on CASSANDRA-14537:
--------------------------------------
Just a brief comment that I think it would be great to think very carefully
about which sub-metrics we expose, as the dev-list email suggests we're
thinking of replicating the entire Codahale/DropWizard API. Since we are going
to force a migration to this new virtual table API, it seems a good opportunity
to isolate the implementation details from our external API, so that we don't
force _another_ migration onto users if we ever overhaul the implementation (or
otherwise unnecessarily constrain the implementation).
min/max, stddev, and perhaps mean, don't make a lot of sense to supply in my
book, and the exponentially weighted moving averages of throughput aren't so
obviously (to me) something we want to maintain or expose for so many
statistics (throughput data is probably being unnecessarily duplicated across
different measures of latency, and we should probably consider exposing them
more explicitly independently of latency numbers). Renaming median to p50, for
consistency with p75+ might also be nice, as examples of discussion points.
> Expose metrics through virtual tables
> -------------------------------------
>
> Key: CASSANDRA-14537
> URL: https://issues.apache.org/jira/browse/CASSANDRA-14537
> Project: Cassandra
> Issue Type: Improvement
> Reporter: Benjamin Lerer
> Assignee: Benjamin Lerer
> Priority: Major
>
> The goal of this ticket is to expose the metrics information through some
> system views (virtual tables).
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
