[
https://issues.apache.org/jira/browse/STORM-671?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14338357#comment-14338357
]
ASF GitHub Bot commented on STORM-671:
--------------------------------------
Github user Lewuathe commented on a diff in the pull request:
https://github.com/apache/storm/pull/442#discussion_r25424494
--- Diff: storm-core/src/clj/backtype/storm/stats.clj ---
@@ -164,13 +164,13 @@
(def COMMON-FIELDS [:emitted :transferred])
(defrecord CommonStats [emitted transferred rate])
-(def BOLT-FIELDS [:acked :failed :process-latencies :executed
:execute-latencies])
+(def BOLT-FIELDS [:acked :failed :process-latencies :executed
:execute-latencies :deserialize-time])
;;acked and failed count individual tuples
-(defrecord BoltExecutorStats [common acked failed process-latencies
executed execute-latencies])
+(defrecord BoltExecutorStats [common acked failed process-latencies
executed execute-latencies deserialize-time])
-(def SPOUT-FIELDS [:acked :failed :complete-latencies])
+(def SPOUT-FIELDS [:acked :failed :complete-latencies :deserialize-time])
--- End diff --
This might be the point I didn't fully understand.
In
[here](https://github.com/apache/storm/blob/master/storm-core/src/clj/backtype/storm/daemon/executor.clj#L413-L429),
all tuple to even spout can go through task-receiver right? I think these
values are sent as bolt ones.
> 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)
