[
https://issues.apache.org/jira/browse/STORM-671?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14481338#comment-14481338
]
ASF GitHub Bot commented on STORM-671:
--------------------------------------
Github user revans2 commented on a diff in the pull request:
https://github.com/apache/storm/pull/442#discussion_r27805711
--- Diff: storm-core/src/storm.thrift ---
@@ -174,12 +174,14 @@ struct BoltStats {
3: required map<string, map<GlobalStreamId, double>> process_ms_avg;
4: required map<string, map<GlobalStreamId, i64>> executed;
5: required map<string, map<GlobalStreamId, double>> execute_ms_avg;
+ 6: required map<string, map<GlobalStreamId, double>> deserialize_time;
}
struct SpoutStats {
1: required map<string, map<string, i64>> acked;
2: required map<string, map<string, i64>> failed;
3: required map<string, map<string, double>> complete_ms_avg;
+ 4: required map<string, map<string, double>> deserialize_time;
--- End diff --
This is still marked as required, if we want a rolling upgrade this needs
to be optional, possibly with a reasonable default, and every time we access
it, mostly for reads, we need to check to see if it is set.
> Measure tuple serialization/deserialization latency.
> ----------------------------------------------------
>
> Key: STORM-671
> URL: https://issues.apache.org/jira/browse/STORM-671
> Project: Apache Storm
> Issue Type: New Feature
> Reporter: Robert Joseph Evans
> Assignee: Kai Sasaki
>
> Some times the serialization/deserialization cost can be very high, and it is
> not currently measured anywhere in storm. We should measure it, at least in
> a similar way to how we do execute and process latency.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
