[
https://issues.apache.org/jira/browse/HBASE-9215?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13740104#comment-13740104
]
Devaraj Das commented on HBASE-9215:
------------------------------------
I am tending to agree with [~eclark]. Since we have metrics for the individual
get/put/etc requests, we probably don't need the FS metrics. I was discussing
with [[email protected]] on whether we should retain these metrics just as
an indicator of overall latencies that we see for FS operations. But that would
be more than just the HStore operations (that the current fs* metrics target)...
> Allow region server metrics to be collected on a periodic basis
> ---------------------------------------------------------------
>
> Key: HBASE-9215
> URL: https://issues.apache.org/jira/browse/HBASE-9215
> Project: HBase
> Issue Type: Improvement
> Reporter: Ted Yu
>
> 0.94 provides the following metrics w.r.t. read / write latencies:
> {code}
> hbase.regionserver.fsReadLatencyHistogram_75th_percentile
> hbase.regionserver.fsReadLatencyHistogram_95th_percentile
> hbase.regionserver.fsReadLatencyHistogram_99th_percentile
> hbase.regionserver.fsReadLatencyHistogram_max
> hbase.regionserver.fsReadLatencyHistogram_mean
> hbase.regionserver.fsReadLatencyHistogram_median
> hbase.regionserver.fsReadLatencyHistogram_min
> hbase.regionserver.fsReadLatencyHistogram_num_ops
> hbase.regionserver.fsReadLatencyHistogram_std_dev
> hbase.regionserver.fsReadLatency_avg_time
> hbase.regionserver.fsReadLatency_num_ops
> hbase.regionserver.fsSyncLatency_avg_time
> hbase.regionserver.fsSyncLatency_num_ops
> hbase.regionserver.fsWriteLatencyHistogram_75th_percentile
> hbase.regionserver.fsWriteLatencyHistogram_95th_percentile
> hbase.regionserver.fsWriteLatencyHistogram_99th_percentile
> hbase.regionserver.fsWriteLatencyHistogram_max
> hbase.regionserver.fsWriteLatencyHistogram_mean
> hbase.regionserver.fsWriteLatencyHistogram_median
> hbase.regionserver.fsWriteLatencyHistogram_min
> hbase.regionserver.fsWriteLatencyHistogram_num_ops
> hbase.regionserver.fsWriteLatencyHistogram_std_dev
> hbase.regionserver.fsWriteLatency_avg_time
> hbase.regionserver.fsWriteLatency_num_ops
> {code}
> The above is collected in doUpdates() which is called periodically.
> In 0.95 / trunk, I don't see facility for periodically updating metrics.
> This should be supported.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
