flink-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Ruby Andrews (JIRA)" <j...@apache.org>
Subject [jira] [Created] (FLINK-10557) Checkpoint size metric incorrectly reports the same value until restart
Date Mon, 15 Oct 2018 21:43:00 GMT
Ruby Andrews created FLINK-10557:
------------------------------------

             Summary: Checkpoint size metric incorrectly reports the same value until restart
                 Key: FLINK-10557
                 URL: https://issues.apache.org/jira/browse/FLINK-10557
             Project: Flink
          Issue Type: Bug
          Components: Metrics
    Affects Versions: 1.4.0
            Reporter: Ruby Andrews


We have seen the following several times, but have not found the root cause. 

The checkpoint size metric will sometimes report the same value over and over, even though
the checkpoint size is changing. The last time we saw this, it happened for 4 days, until
we re-started the Flink cluster. In that time period, the application flushes all data each
day so we would expect to see the checkpoint size grow until UTC midnights, then go to about
0 and begin growing again.

It appears that the metrics continue to be gathered, because we see them in our data repository
where we are reporting them. However, the size does not change.  

Is there more information we can gather to root cause this if it happens again?

 

 



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Mime
View raw message