Filed from: Test Regression Details

Test Regression:

[sig-network-edge][Feature:Idling] Idling with a single service and ReplicationController should idle the service and ReplicationController properly [Suite:openshift/conformance/parallel]

This test, related to the Networking / router component and Idling capability, has shown a significant regression. The regression was opened on October 9, 2025 (5 days ago) and is still ongoing. The status code -400 indicates a "SignificantRegression detected", with a 100.00% Fisher's Exact probability of regression. The test pass rate has dropped from 99.52% to 88.00%.

Statistics

Sample (being evaluated) Release: 4.20
Start Time: 2025-10-06
End Time: 2025-10-13
Success Rate: 88.0%
Successes: 20
Failures: 3
Flakes: 2

Base (historical) Release: 4.19
Start Time: 2025-05-18
End Time: 2025-06-17
Success Rate: 99.52%
Successes: 199
Failures: 1
Flakes: 9

Sample Failure Outputs

Job Run 1976162309969547264:
fail [github.com/openshift/origin/test/extended/idling/idling.go:129]: Failed after 2.911s.
Expected
<string>: "2"
to contain substring
<string>: 0

Job Run 1976246640712880128:
fail [github.com/openshift/origin/test/extended/idling/idling.go:129]: Failed after 3.387s.
Expected
<string>: "2"
to contain substring
<string>: 0

Job Run 1975680050564960256:
fail [github.com/openshift/origin/test/extended/idling/idling.go:129]: Failed after 3.476s.
Expected
<string>: "2"
to contain substring
<string>: 0

Links to Relevant Jobs

• periodic-ci-openshift-release-master-nightly-4.20-e2e-vsphere-ovn-techpreview (1976162309969547264)
• periodic-ci-openshift-release-master-nightly-4.20-e2e-vsphere-ovn-techpreview (1976246640712880128)
• periodic-ci-openshift-release-master-nightly-4.20-e2e-vsphere-ovn-disk-setup-techpreview (1975680050564960256)

Patterns and Insights

The test

[sig-network-edge][Feature:Idling] Idling with a single service and ReplicationController should idle the service and ReplicationController properly [Suite:openshift/conformance/parallel]

is consistently failing with the same error message across multiple job runs: "Expected <string>: \"2\" to contain substring <string>: 0". This indicates a specific issue in the idling logic or the test's expectation. The base release (4.19) had a very high success rate with some flakes, while the sample release (4.20) shows a significant increase in failures and a decrease in flakes, suggesting a change that converted previous flaking behavior into outright failures. This regression represents a critical quality concern and should be treated as a release blocker.

Filed by: lmeyer@redhat.com