[
https://issues.apache.org/jira/browse/HBASE-3102?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12925144#action_12925144
]
HBase Review Board commented on HBASE-3102:
-------------------------------------------
Message from: "Gary Helmling" <[email protected]>
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://review.cloudera.org/r/1055/#review1672
-----------------------------------------------------------
Looks pretty good. These look like great metrics to add!
Only issue I see is some odd naming that will happen for attributes used for
the "*Size" metrics. I'd consider copying the rest of the implementation from
MetricsTimeVaryingRate -> PersistentMetricsTimeVaryingRate and extending
MetricsBase directly, to allow tweaking the naming. Thoughts?
A few trailing whitespace issues throughout. Otherwise looks great!
trunk/src/main/java/org/apache/hadoop/hbase/master/metrics/MasterMetrics.java
<http://review.cloudera.org/r/1055/#comment5547>
I think using this type will result in the average split size for the
period being reported as "splitSize_avg_time"? The value should be right, the
attribute name is just a little weird since it's size not time.
trunk/src/main/java/org/apache/hadoop/hbase/master/metrics/MasterMetrics.java
<http://review.cloudera.org/r/1055/#comment5546>
Drop trailing whitespace here and following sections
trunk/src/main/java/org/apache/hadoop/hbase/metrics/PersistentMetricsTimeVaryingRate.java
<http://review.cloudera.org/r/1055/#comment5548>
Given the above naming of "splitSize_avg_time", maybe consider extending
MetricsBase directly and essentially copy-n-paste forking
MetricsTimeVaryingRate?
I know it sucks not being able to subclass cleanly here, but then you could
at least expose the attributes as:
- name + "_num_ops"
- name + "_avg"
and avoid the naming confusion.
trunk/src/main/java/org/apache/hadoop/hbase/regionserver/metrics/RegionServerMetrics.java
<http://review.cloudera.org/r/1055/#comment5549>
Same slightly weird naming for "compactionSize" and "flushSize" metrics
will happen as for "splitSize" in MasterMetrics.
- Gary
> Enhance HBase rMetrics for Long-running Stats
> ---------------------------------------------
>
> Key: HBASE-3102
> URL: https://issues.apache.org/jira/browse/HBASE-3102
> Project: HBase
> Issue Type: Improvement
> Components: regionserver
> Reporter: Nicolas Spiegelberg
> Assignee: Nicolas Spiegelberg
> Fix For: 0.90.0
>
> Attachments: HBASE-3102_90.patch
>
>
> Example: a useful metric to observe track is compaction count + duration.
> Since compactions are long running and only happen infrequently, the
> avg/opcount stats are should be reset longer than the default polling period
> (5 sec). In addition to 'hbase.period', we should allow a different duration
> after which long-running metrics should expire. This would also fix our
> existing metrics problem where min/max stats are never reset until the
> process is restarted/upgraded.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
