e2e: Re-enable and fix flaky pipelinerun-with-matrix example test

[vdemeester]

Problem

The TestExamples/v1/pipelineruns/beta/pipelinerun-with-matrix e2e test has been consistently flaky in CI, causing build failures. The test file has been temporarily moved to the no-ci folder to prevent CI failures while we investigate and fix
the root cause.

The test was disabled in PR #9200 as a temporary measure.

Evidence of Flakiness

From issue #9062, the test has been failing with timeouts:

--- FAIL: TestExamples/v1/pipelineruns/beta/pipelinerun-with-matrix (605.80s)

Recent CI failures:

• https://github.com/tektoncd/pipeline/actions/runs/18324465586/job/52185466094?pr=9047
• https://github.com/tektoncd/pipeline/actions/runs/18324465586/job/52185466098?pr=9047
• https://github.com/tektoncd/pipeline/actions/runs/18339183125/job/52230530489?pr=9067

Tasks

• Investigate why the pipelinerun-with-matrix example test is timing out (600+ seconds)
• Identify the root cause of the flakiness
• Implement a fix for the underlying issue
• Move the test file back from examples/v1/pipelineruns/no-ci/pipelinerun-with-matrix.yaml to examples/v1/pipelineruns/beta/pipelinerun-with-matrix.yaml
• Verify the test runs reliably in CI

Current Location

The test file is currently at: examples/v1/pipelineruns/no-ci/pipelinerun-with-matrix.yaml

It should be moved back to: examples/v1/pipelineruns/beta/pipelinerun-with-matrix.yaml once fixed.

Related Issues

• e2e: TestExamples is flaky #9062 - Parent issue tracking flaky e2e example tests

[vdemeester]

Update: Extended to Include Second Failing Test

After investigating recent CI failures, identified two related flaky tests that need to be fixed:

1. pipelinerun-with-matrix

• Location: examples/v1/pipelineruns/beta/pipelinerun-with-matrix.yaml
• Failures:
  • Run 20265445080 (Dec 16): 902.94s timeout
  • Run 20242616379 (Dec 15): 903.07s timeout
  • Run 20231100917 (Dec 15): 903.02s timeout
• Pattern: Fails on k8s-oldest + alpha, passes on k8s-latest
• Suspect: matrix-with-task-retries task with complex retry logic

2. using-retries-and-retry-count-variables

• Location: examples/v1/pipelineruns/using-retries-and-retry-count-variables.yaml
• Failures:
  • Run 20276197820 (Dec 16): 900.27s timeout
• Pattern: Same as above - k8s-oldest + alpha only
• Suspect: Similar retry logic with context variables

Common Root Cause

Both tests timeout at exactly 15 minutes (global test timeout) on k8s-oldest with alpha features. Hypothesis:

• Context variables like $(context.pipelineTask.retries) or $(context.task.retry-count) may not resolve correctly on older Kubernetes versions with alpha features
• This causes infinite retry loops until timeout

PR Status

✅ PR #9242 - Both tests moved to no-ci/ temporarily to unblock CI

Investigation Plan

Created detailed analysis note with:

• Reproduction steps (kind cluster with k8s 1.28.0 + alpha features)
• Minimal reproducer creation
• Three proposed fix options

Next: Reproduce locally and identify exact root cause.