[
https://issues.apache.org/jira/browse/CASSANDRA-10636?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14986958#comment-14986958
]
Stefan Podkowinski commented on CASSANDRA-10636:
------------------------------------------------
I've recently created a trivial patch to monitor the execution time for each
native transport message type. It's not specific to CQL parsing, but can also
be used to derive the time spend between receiving the message from netty and
StoryProxy execution.
> Add metrics to detect time spent in message processing and parsing
> -------------------------------------------------------------------
>
> Key: CASSANDRA-10636
> URL: https://issues.apache.org/jira/browse/CASSANDRA-10636
> Project: Cassandra
> Issue Type: Improvement
> Reporter: Soumava Ghosh
>
> We use StorageProxy latencies to get an idea of the latency from Cassandra.
> However, all of the transport layer including CQL parsing happens before this
> so we don't have visibility into how long these are taking. We should record
> latencies earlier on in the processing and provide a view of the total time
> taken inside Cassandra.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
