gnodet commented on PR #2971:
URL: https://github.com/apache/cxf/pull/2971#issuecomment-4134669691

   _Claude Code on behalf of Guillaume Nodet_
   
   I analyzed recent CI builds (last month) to check how frequently this 
timeout issue occurs. The results show it's pervasive:
   
   ### Builds with tracing `ConditionTimeoutException` at 5s
   
   | Date | Run | Branch | Failing Tests |
   |------|-----|--------|---------------|
   | Mar 25 | `23522519679` | `rat-plugin-0.18` | Micrometer, OTel, Brave (JDK 
17 & 21) |
   | Mar 25 | `23522374425` | `swagger-ui-5.32.1` | Micrometer, OTel |
   | Mar 20 | `23327005846` | `atmosphere-4.0.14` | Micrometer, OTel |
   | Mar 17 | `23176075016` | `atmosphere-4.0.11` | Micrometer, Brave |
   | Mar 16 | `23142577728` | `fix.collissions` | Micrometer, OTel |
   | Mar 16 | `23124650038` | `fix.collissions` | Micrometer, Brave |
   | Mar 16 | `23123226388` | `fix.collissions` | Micrometer |
   | Mar 15 | `23118883527` | `fix.collissions` | Micrometer, OTel |
   | Mar 13 | `23072188687` | `mtom-chunking` | Micrometer, OTel |
   | Mar 13 | `23070574942` | `oauth2-response-leak` | Micrometer, OTel |
   | Mar 13 | `23054511418` | `oauth2-response-leak` | Micrometer, OTel |
   
   The failure appears in nearly every build, across multiple branches and JDK 
versions (17 & 21). Tests time out at exactly ~5.15s, confirming it's the 
Awaitility `atMost(5s)` ceiling.
   
   ### 10s is also not enough
   
   Run `23065827441` (`cleanup/remove-pre-servlet31-compat`, Mar 13) shows 
`OpenTelemetryTracingTest.testThatNewSpanIsCreatedOnClientTimeout` failing at 
**10 seconds**:
   ```
   ConditionTimeoutException: Condition with Lambda expression in
   org.apache.cxf.systest.jaxrs.tracing.opentelemetry.OpenTelemetryTracingTest
   was not fulfilled within 10 seconds.
   ```
   
   ### Even "successful" builds are affected
   
   Builds like the Mar 25 dependabot PRs (`23522519679`, `23522374425`) are 
marked SUCCESS despite having tracing test failures — the CI uses `mvn -fae` so 
the tracing module failures don't block the overall build. The failures are 
silently ignored.
   
   ### Updated to 30s
   
   Pushed an update bumping from 10s to 30s to provide sufficient headroom for 
loaded CI environments. The timeout is purely a safety net — the conditions 
normally resolve in well under 1 second, so raising it doesn't slow down 
passing tests.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to