[
https://issues.apache.org/jira/browse/CASSANDRA-11852?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15293509#comment-15293509
]
Chris Lohfink edited comment on CASSANDRA-11852 at 5/20/16 3:09 PM:
--------------------------------------------------------------------
The table metrics do this in
{{org.apache.cassandra.metrics:type=ColumnFamily,name=...}} or
{{org.apache.cassandra.metrics:type=Table,name=...}} (newer versions) for
metric that can be tracked per table easily.
ie:
{code}org.apache.cassandra.metrics:type=ColumnFamily,keyspace=system,scope=peers,name=ReadLatency
{code}
Read repair is one that could probably be tracked per table but is tracked
globally. but for most part the global ones are that way due to some
complication or oversight in the implementation.
was (Author: cnlwsu):
The table metrics do this in
{{org.apache.cassandra.metrics:type=ColumnFamily,name=...}} or
{{org.apache.cassandra.metrics:type=Table,name=...}} (newer versions) for
metric that can be tracked per table easily. Read repair is one that could
probably be tracked per table but is tracked globally. but for most part the
global ones are that way due to some complication or oversight in the
implementation.
> Feature request: Cluster metrics should include keyspace and tablename where
> possible.
> --------------------------------------------------------------------------------------
>
> Key: CASSANDRA-11852
> URL: https://issues.apache.org/jira/browse/CASSANDRA-11852
> Project: Cassandra
> Issue Type: Improvement
> Reporter: Sucwinder Bassi
>
> For example, for read repair metrics
> Metric name (from cluster):
> llds.cassandra.llds_1_acc.cluster.dcr.lrv140gh.org.apache.cassandra.metrics.ReadRepair.RepairedBlocking.count
> Proposed (added <Keyspace Name>.<Table Name>):
> llds.cassandra.llds_1_acc.cluster.dcr.lrv140gh.org.apache.cassandra.metrics.ReadRepair.<Keyspace
> Name>.<Table Name>.RepairedBlocking.count
> It would be beneficial to add <Keyspace Name>.<Table Name> where possible in
> the metrics as this will help narrow down the investigation and remove guess
> work as to which object is being impacted.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
