[
https://issues.apache.org/jira/browse/CASSANDRA-12649?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15665310#comment-15665310
]
Alwyn Davis commented on CASSANDRA-12649:
-----------------------------------------
I've updated the patch - [^12649-3.x-v2.patch] based on your points.
I've also added a {{BatchBench}} JMH class and ran it against trunk, the first
implementation of this patch and this current implementation:
*BATCH METRICS*
{noformat}o.a.c.t.m.BatchBench.batchLogged thrpt 50 38.784
± 6.790 ops/ms
o.a.c.t.m.BatchBench.batchUnlogged thrpt 50 39.278 ± 7.572
ops/ms{noformat}
*BATCH METRICS (v2)*
{noformat}o.a.c.t.m.BatchBench.batchLogged thrpt 50 40.340
± 7.856 ops/ms
o.a.c.t.m.BatchBench.batchUnlogged thrpt 50 38.198 ± 8.116
ops/ms{noformat}
*TRUNK*
{noformat}o.a.c.t.m.BatchBench.batchLogged thrpt 50 38.737
± 7.661 ops/ms
o.a.c.t.m.BatchBench.batchUnlogged thrpt 50 38.455 ± 7.111
ops/ms{noformat}
As {{update.dataSize()}} is not checked on single partition batches,
{{mutationSizeHistogram}} will only measure multi-partition batches.
> Add BATCH metrics
> -----------------
>
> Key: CASSANDRA-12649
> URL: https://issues.apache.org/jira/browse/CASSANDRA-12649
> Project: Cassandra
> Issue Type: Wish
> Reporter: Alwyn Davis
> Priority: Minor
> Fix For: 3.x
>
> Attachments: 12649-3.x-v2.patch, 12649-3.x.patch,
> stress-batch-metrics.tar.gz, stress-trunk.tar.gz, trunk-12649.txt
>
>
> To identify causes of load on a cluster, it would be useful to have some
> additional metrics:
> * *Mutation size distribution:* I believe this would be relevant when
> tracking the performance of unlogged batches.
> * *Logged / Unlogged Partitions per batch distribution:* This would also give
> a count of batch types processed. Multiple distinct tables in batch would
> just be considered as separate partitions.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
