[
https://issues.apache.org/jira/browse/METRON-1060?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16099883#comment-16099883
]
ASF GitHub Bot commented on METRON-1060:
----------------------------------------
Github user justinleet commented on a diff in the pull request:
https://github.com/apache/metron/pull/665#discussion_r129279181
--- Diff:
metron-platform/metron-enrichment/src/main/java/org/apache/metron/enrichment/bolt/GenericEnrichmentBolt.java
---
@@ -179,6 +180,7 @@ public void declareOutputFields(OutputFieldsDeclarer
declarer) {
@SuppressWarnings("unchecked")
@Override
public void execute(Tuple tuple) {
+ long texecute1 = System.currentTimeMillis();
--- End diff --
I don't know if we actually care given that it's logging, but iirc
`currentTimeMillis()` returns can differ based on ntp syncs and so on (end time
can be before start time, for example). `nanotime()` is generally more
reliable for this, but it can also be more expensive (on the order of millis).
Even with `currentTimeMillis()`, I think the actual call can still end up
being slow (milli+ depending on the system). It might be worthwhile to wrap
the timing calls in a debug guard given that these run in the `execute()`, but
I'm not sure if it'll make a difference in practice.
> Add performance timing logging to enrichment topology
> -----------------------------------------------------
>
> Key: METRON-1060
> URL: https://issues.apache.org/jira/browse/METRON-1060
> Project: Metron
> Issue Type: Improvement
> Reporter: Michael Miklavcic
> Assignee: Michael Miklavcic
>
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
