[
https://issues.apache.org/jira/browse/TEZ-4589?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
László Bodor updated TEZ-4589:
------------------------------
Description:
We need to track the number of failed task attempts, otherwise, there is a
chance they remain "invisible" in case of a successful DAG with performance
degradation. From the counters, ideally, we should be able to the the overall
time spent in the FAILED attempts.
UPDATE: we already have NUM_FAILED_TASKS, which might be misleading, as it's
just attempts, anyway, still need to aggregate the time
was:We need to track the number of failed task attempts, otherwise, there is
a chance they remain "invisible" in case of a successful DAG with performance
degradation. From the counters, ideally, we should be able to the the overall
time spent in the FAILED attempts.
> Counter for the overall duration of failed task attempts
> --------------------------------------------------------
>
> Key: TEZ-4589
> URL: https://issues.apache.org/jira/browse/TEZ-4589
> Project: Apache Tez
> Issue Type: Improvement
> Reporter: László Bodor
> Assignee: László Bodor
> Priority: Major
>
> We need to track the number of failed task attempts, otherwise, there is a
> chance they remain "invisible" in case of a successful DAG with performance
> degradation. From the counters, ideally, we should be able to the the overall
> time spent in the FAILED attempts.
> UPDATE: we already have NUM_FAILED_TASKS, which might be misleading, as it's
> just attempts, anyway, still need to aggregate the time
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
