Repository navigation
[JENKINS-57888] Custom log format broken by 2.177 #20454
Description
Upon upgrading from 2.176 to 2.177, it would appear that support for providing a custom log format has been broken, due to what looks like an undocumented change (jenkinsci/winstone#63) in the Winstone library.
Logging.properties file:
.level=ALL handlers=java.util.logging.ConsoleHandler java.util.logging.SimpleFormatter.format=[%1$tF %1$tT.%1$tL][%4$s][%2$s] %5$s %6$s%n java.util.logging.ConsoleHandler.level=INFO java.util.logging.ConsoleHandler.formatter=java.util.logging.SimpleFormatter
I've also added the following JVM argument:
-Djava.util.logging.config.file=/var/lib/jenkins/logging.properties
Log format with 2.176:
[2019-06-06 15:10:07.881][INFO][jenkins.InitReactorRunner$1 onAttained] Completed initialization
[2019-06-06 15:10:08.052][INFO][hudson.WebAppMain$3 run] Jenkins is fully up and running
Log format with 2.177:
2019-06-06 15:11:40.956+0000 [id=29] INFO jenkins.InitReactorRunner$1#onAttained: Completed initialization
2019-06-06 15:11:41.136+0000 [id=20] INFO hudson.WebAppMain$3#run: Jenkins is fully up and running
Originally reported by 
fatmcgav, imported from: Custom log format broken by 2.177
- assignee: l_r
- status: Resolved
- priority: Major
- component(s): core, winstone-jetty
- label(s): 2.204.3-fixed, 2.204.5-rejected, 2.222.1-fixed, non-trivial-lts-backporting, regression
- resolution: Fixed
- resolved: 2020-01-27T22:43:37+00:00
- votes: 3
- watchers: 8
- imported: 2025-11-24
Raw content of original issue
Upon upgrading from 2.176 to 2.177, it would appear that support for providing a custom log format has been broken, due to what looks like an undocumented change (jenkinsci/winstone#63) in the Winstone library.
Logging.properties file:
.level=ALL handlers=java.util.logging.ConsoleHandler java.util.logging.SimpleFormatter.format=[%1$tF %1$tT.%1$tL][%4$s][%2$s] %5$s %6$s%n java.util.logging.ConsoleHandler.level=INFO java.util.logging.ConsoleHandler.formatter=java.util.logging.SimpleFormatterI've also added the following JVM argument:
-Djava.util.logging.config.file=/var/lib/jenkins/logging.properties
Log format with 2.176:
[2019-06-06 15:10:07.881][INFO][jenkins.InitReactorRunner$1 onAttained] Completed initialization [2019-06-06 15:10:08.052][INFO][hudson.WebAppMain$3 run] Jenkins is fully up and running
Log format with 2.177:
2019-06-06 15:11:40.956+0000 [id=29] INFO jenkins.InitReactorRunner$1#onAttained: Completed initialization 2019-06-06 15:11:41.136+0000 [id=20] INFO hudson.WebAppMain$3#run: Jenkins is fully up and running
Activity
jenkins-infra-bot commentedon Jun 6, 2019
Remote link associated with this issue:
jenkins-infra-bot commentedon Jan 15, 2020
l_r:
Raw content of original comment:
Is there a workaround to be able to provide a ConsoleHandler formatter?
If not, I'll be submitting a PR removing the overriding of it.
It will still be possible to set the formatter from the command line or the logging.properties as usual.
What do you think?
jenkins-infra-bot commentedon Jan 15, 2020
Raw content of original comment:
The new formatter is deliberate. There could be a Winstone option to select an alternate formatter if someone cares to write it, or perhaps just by disabling the new formatter when java.util.logging.config.file is defined.
jenkins-infra-bot commentedon Jan 16, 2020
l_r:
Raw content of original comment:
Yes, that would be possible.
I'm wondering, though, why the standard java behavior is overridden. It's not something one would expect.
jenkins-infra-bot commentedon Jan 24, 2020
Raw content of original comment:
#4452
jenkins-infra-bot commentedon Jan 27, 2020
Raw content of original comment:
It was released in 2.218, thanks to l_r! I will mark it as LTS candidate, but it would require Winstone backporting due to the
JENKINS-60857regressions in 5.6. If olivergondza decides it is important enough, I will do such backporting.jenkins-infra-bot commentedon Mar 6, 2020
Raw content of original comment:
This will be reverted in 2.204.5 as it has caused regressions much more impactfull.
jenkins-infra-bot commentedon Mar 6, 2020
Raw content of original comment:
To be clear, it was not this fix which caused the regressions, but unrelated changes in winstone.