Flaky E2E: paginated exporter listing fails with etcd request timeout

[raballew]

Summary

The E2E test "paginated exporter listing returns all exporters" (e2e/test/e2e_test.go:343) is flaky. It fails intermittently with an etcd timeout when creating many exporters for pagination testing.

Failure Details

• Test: Core E2E Tests > Lease operations > paginated exporter listing returns all exporters
• File: e2e/test/e2e_test.go:343
• Duration before failure: 2m52s (likely hit a timeout)
• Root cause: Kubernetes API returns HTTP 500 while creating exporter pagination-exp-93:

  ApiException: (500)
  Reason: Internal Server Error
  HTTP response body: {"kind":"Status","apiVersion":"v1","metadata":{},"status":"Failure","message":"etcdserver: request timed out","code":500}
  

Observed In

• PR feat: shell completion for jmp, jmp-admin, and j #422 (039-shell-completion) - CI run
• PR feat: shell completion for jmp, jmp-admin, and j #422 only changes Python CLI completion files and a test workflow YAML, with zero changes to Go code, E2E tests, or controller logic. The failure is unrelated to the PR.

Analysis

The test creates ~100 exporters sequentially to test pagination. Under CI resource constraints, the kind cluster's etcd can timeout around exporter 93 of 100. This is an infrastructure/resource pressure issue, not a code bug.

Possible Mitigations

• Reduce the number of exporters created (e.g., from 100 to 25, with a smaller page size)
• Add retry logic around exporter creation in the test
• Add a brief sleep between creation batches to reduce etcd pressure
• Create exporters in parallel with a semaphore to control concurrency