Failing test: Workflow access dialog - lets an executor test a saved workflow with enabled=false

A test failed on a tracked branch

Scout Test Details:

Field	Value
Test ID	478a0cfe64d4fb9-44fe125f3b43fe5
Target	cloud-stateful-classic
Location	src/platform/plugins/shared/workflows_management/test/scout/ui/parallel_tests/workflow_access_control.spec.ts
Duration	34.23s
Module	workflowsManagement (plugin)
Config path	src/platform/plugins/shared/workflows_management/test/scout/ui/parallel.playwright.config.ts
Code Owners	elastic/workflows-eng

TimeoutError: locator.click: Timeout 10000ms exceeded.
Call log:
  - waiting for getByRole('option', { name: 'test editor' })

    at Proxy.addAccessUser (src/platform/plugins/shared/workflows_management/test/scout/ui/fixtures/page_objects/workflow_editor_page.ts:74:11)
    at src/platform/plugins/shared/workflows_management/test/scout/ui/parallel_tests/workflow_access_control.spec.ts:112:11

First failure: appex-qa-stateful-kibana-scout-tests - main

Failure screenshots are available in the Buildkite HTML report and artifacts.

New failure for "cloud-stateful-classic" target: appex-qa-stateful-kibana-scout-tests - main

Error message matches a failure already reported on this issue.

✅ Flaky Test Fixer completed successfully!

Generated by Flaky Test Fixer for #294772

➡️ A fix PR is ready for review: #294917

The access-control spec now resolves each participant's display name and username from /internal/security/me at runtime instead of hard-coding the local-only SAML identities, which never exist on Cloud.

Generated by Flaky Test Fixer for #294772 · claude · opus · 356.9 AIC · ⌖ 61.8 AIC · ⊞ 15K · ◷

New failure for "cloud-stateful-classic" target: appex-qa-stateful-kibana-scout-tests - main

Error message matches a failure already reported on this issue.

New failure for "cloud-stateful-classic" target: appex-qa-stateful-kibana-scout-tests - main

Error message matches a failure already reported on this issue.

New failure for "cloud-stateful-classic" target: appex-qa-stateful-kibana-scout-tests - main

Error message matches a failure already reported on this issue.

CI environment issue — the failing build again predates the merged fix

Build 561 ran Kibana e99c7c7 (5 Oct 10:36 UTC), which is 16 commits behind the fix a0c6268 (5 Oct 12:06 UTC), so addAccessUser('test editor') still timed out on a literal that no longer exists on main.

Details

Proposed fix

No change for this test — #294917 already fixed it and the Cloud image simply hasn't picked it up yet.

One follow-up is still needed, though, and it is not covered by that PR: workflow_access_control_admin.spec.ts#L69-L70 still hard-codes addAccessUser('test admin') / setAccessRole('elastic_admin', 'executor'). Lift the resolveUser helper from workflow_access_control.spec.ts#L18-L21 into a shared module and use it for the admin role there too. Unlike this issue's test, that one will keep failing on Cloud after the image catches up — same Design tests with a cloud-first mindset rule.

Root cause & evidence

The build does not contain the fix. GET /repos/elastic/kibana/compare/a0c6268...e99c7c7 returns status: "behind", behind_by: 16, and merge_base_commit.sha == e99c7c7 — i.e. the checkout is an ancestor of the fix. The job log confirms the pin: Build hash: e99c7c795f342e2a221499dc752e166a43e41ae4 in the Deployment information block, then Checking out commit 'e99c7c79...' for /root/.qaf/data/git/kibana, the directory every failing stack frame is rooted in.
The signature is byte-for-byte pre-fix, including the old line numbers: saves private and public access through the shared form fails at spec.ts:69:10 on getByText('test editor', { exact: true }), and both executor cases at spec.ts:119:13 on getByRole('option', { name: 'test editor' }). Post-fix those literals are gone — the only surviving match for test editor in the spec is the explanatory comment above resolveUser.
Not failure:fix-did-not-hold: the fix was merged 5 Oct 12:06 UTC and this build's checkout is from 5 Oct 10:36 UTC — 90 minutes earlier. The fix has still never been exercised on this pipeline.
Correction to the previous verdict: it claimed main "no longer contains any of the local-only identities". That is true for this spec but not for the sibling admin spec, which failed in the same job on getByRole('option', { name: 'test admin' }) (workflow_access_control_admin.spec.ts:56:5) and is genuinely unfixed on main. No open PR addresses it.

Additional context

Ruled out: a new or different flake — all three cases of this spec fail on the pre-fix hard-coded strings at pre-fix line offsets, and the two browser 404s logged at 12:50:21Z / 12:50:33Z bracket the click window but are incidental (the _suggest_user_profiles call succeeded, as the previous investigation established).
Verification: for the next reported failure, resolve the run's Build hash from the Deployment information block and run gh api repos/elastic/kibana/compare/a0c626839bd5456a8a8ae33b8913858716f4aff5...<build-hash> --jq '.status' — ahead/identical means the fix was genuinely exercised and the failure counts.

Note

Closing this issue: the failing run used Kibana e99c7c7, which predates the fix commit a0c6268 by 90 minutes, so it ran pre-fix code rather than a recurrence. It will reopen automatically if the test fails again. The sibling admin-spec gap noted above needs its own fix and is not tracked by this issue.

Generated by Failed Test Investigator for #294772 · claude · opus · 256.3 AIC · ⌖ 55 AIC · ⊞ 14.1K · ◷

Closing: the failing run (build 561) used Kibana e99c7c795f342e2a221499dc752e166a43e41ae4 (5 Oct 10:36 UTC), which is 16 commits behind the fix commit a0c626839bd5456a8a8ae33b8913858716f4aff5 (5 Oct 12:06 UTC), so it ran pre-fix code — propagation lag, not a recurrence. See the investigation comment above. This will reopen automatically if the test fails again.

Note: the sibling workflow_access_control_admin.spec.ts has the same local-only SAML identities still hard-coded on main and needs a separate fix; that is not tracked by this issue.

Generated by Failed Test Investigator for #294772 · claude · opus · 256.3 AIC · ⌖ 55 AIC · ⊞ 14.1K

New failure for "cloud-stateful-classic" target: appex-qa-stateful-kibana-scout-tests - main

Error message matches a failure already reported on this issue.

Closing: build 562 ran Kibana e99c7c795f342e2a221499dc752e166a43e41ae4 — the same pre-fix checkout as build 561 — so it ran pre-fix code. Propagation lag, not a recurrence. Verdict and root cause are unchanged from the previous investigation, so no new analysis comment.

Confirmed directly from the source at that commit rather than by date arithmetic: workflow_access_control.spec.ts @ e99c7c7 still hard-codes getByText('test editor', { exact: true }) (L75), addAccessUser('test editor'), 'test viewer', 'elastic_viewer' and 'elastic_editor' — all removed by #294917 (merged 5 Oct 12:42 UTC), which main now carries via resolveUser / samlAuth.session.getUserData. The job's code frame matches that checkout line-for-line (67 saveWorkflow → 69 if (!workflowId ...)), confirming the pin; the Deployment information block reports Build hash: e99c7c79... and the run then checks that commit out into /root/.qaf/data/git/kibana, the directory every failing stack frame is rooted in.

Two notes for whoever sees this next:

The Cloud image has not advanced in ~2 days. Builds 561 (6 Oct) and 562 (7 Oct) both pinned e99c7c79 (5 Oct 10:36 UTC), ~38 h behind the fix merge. Expect this issue to reopen and re-close on each nightly run until the image moves past a0c6268; that is image-promotion lag, not a test problem. The fix has still never been exercised on this pipeline.
The sibling admin spec is genuinely unfixed on main and failed in the same job: workflow_access_control_admin.spec.ts#L69-L70 still hard-codes addAccessUser('test admin') / setAccessRole('elastic_admin', 'executor'). Unlike this issue's spec it will keep failing on Cloud after the image catches up. No open PR addresses it, and it is not tracked here — it needs its own issue.

To check the next reported failure: resolve the run's Build hash from the Deployment information block and run gh api repos/elastic/kibana/compare/a0c626839bd5456a8a8ae33b8913858716f4aff5...<build-hash> --jq '.status' — ahead/identical means the fix was genuinely exercised and the failure counts. This will reopen automatically if the test fails again.

Generated by Failed Test Investigator for #294772 · claude · opus · 222.4 AIC · ⌖ 39.5 AIC · ⊞ 14.1K

Navigation Menu

Repository navigation

Failing test: Workflow access dialog - lets an executor test a saved workflow with enabled=false #294772

➡️ A fix PR is ready for review: #294917

9 remaining items

CI environment issue — the failing build again predates the merged fix

Proposed fix

Root cause & evidence

Additional context

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Failing test: Workflow access dialog - lets an executor test a saved workflow with enabled=false #294772

Description

Activity

github-actions commented on Oct 1, 2026

kibanamachine commented on Oct 2, 2026

github-actions commented on Oct 2, 2026

github-actions commented on Oct 2, 2026

➡️ A fix PR is ready for review: #294917

kibanamachine commented on Oct 2, 2026

kibanamachine commented on Oct 3, 2026

9 remaining items

github-actions commented on Oct 5, 2026

github-actions commented on Oct 5, 2026

kibanamachine commented on Oct 6, 2026

github-actions commented on Oct 6, 2026

CI environment issue — the failing build again predates the merged fix

Proposed fix

Root cause & evidence

Additional context

github-actions commented on Oct 6, 2026

kibanamachine commented on Oct 7, 2026

github-actions commented on Oct 7, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions