[Rohit0301]

Describe your changes:

Fixes the merge-queue flakes reported by the Playwright Flaky Daily Report run 40998300289:

• Features/AIMode/CustomPropertiesPanel.spec.ts › edits a custom property display name and description
• Pages/AIProfile/ProfileTeam.spec.ts › Should search teams in the Organization sub-teams table
• Features/AIMode/PersonaSettings.spec.ts › Home page: a configured Curated Assets widget is saved

Why they started on 2026-10-29: none of these tests appears in any earlier daily report, and neither the tests nor their features changed. The CI shard planner packs specs by rolling timing history, and #35902 reshaped its lanes. After the repack these tests ran next to new neighbours under more load. For example, CustomPropertiesPanel moved into the same shard as Pages/CustomProperties.spec.ts. That exposed timing races in the tests. Evidence comes from the CI blob reports' failure snapshots.

Test	Root cause	Fix (test-only)
CustomPropertiesPanel edit	The save PATCH returned 200, but the row still showed Initial description. Parallel specs, and sibling tests in this spec, were adding and removing properties on the shared table type in the same second and overwrote the edit.	The edit test now uses the worksheet type. No other spec adds, edits or removes worksheet custom properties; only entity-data-setup touches it, before the tests run. The helpers take the type as a parameter and default to table.
ProfileTeam search	The search box waits briefly after typing before applying the search, and applying it writes the term back into the box. The test only checked that the team was visible, which was already true before any filter ran, so it typed `non-existent-team-xyz` before the first search applied. The late write-back then reverted the box, as the CI snapshot shows, and the team stayed visible.	The search filters the already-loaded teams in the browser, so there is no request to await; the table state is the signal. Each step now waits for its result before the next: search, then only that team listed; clear, then the box empty and the full list back; new term, then the box holds it and the team is hidden. No retry loop.
PersonaSettings curated widget	(a) A cold load of `/#personas/...` raced the boot's auth and landing redirects, which can replace the URL before the settings store reads its hash. The app settled on Home and `GET /personas/name/...` never fired (3-minute timeout). (b) Picking Table fires a debounced asset count, and its banner renders inside the dialog. A local failure trace shows the count landing between opening the operator list and the click, so the shifted list picked "Not in" and Save stayed disabled.	(a) `openPersonaSettings` boots through `redirectToHomePage` (which already handles first-boot recovery) and then opens the deep link inside the running app. The custom-properties wait now counts only requests issued after it is armed, so a fetch the Home page still has in flight cannot satisfy it. (b) The test waits for that count request and its banner before building the rule, then picks field, operator and value with web-first assertions. No retry loops.

Type of change:

• Bug fix

High-level design:

N/A — test-only change to three specs.

Tests:

Use cases covered

• Edit a custom property's display name and description from the AI-mode Custom Properties panel
• Search and clear-search the Organization sub-teams table
• Configure and save a Curated Assets widget on a persona's Home page layout via its deep link

Unit tests

Not applicable (no app code changes).

Backend integration tests

• Not applicable (no backend API changes).

Ingestion integration tests

• Not applicable (no ingestion changes).

Playwright (UI) tests

• Files updated: CustomPropertiesPanel.spec.ts, ProfileTeam.spec.ts, PersonaSettings.spec.ts
• Ran the three tests 10× each with 3 workers, no retries and `--no-deps` (`entity-data-setup` not run), against a local server: 30 passed (3.6m). This run used the first version of the PersonaSettings fix (first screenshot).
• After replacing the PersonaSettings retries with the deterministic waits above:
  • Curated Assets test 10× with 3 workers: 10 passed (2.4m).
  • Whole `PersonaSettings.spec.ts` once with 3 workers, since every test uses `openPersonaSettings`: 20 passed (2.0m) (second screenshot).
• After the ProfileTeam search → clear → search change above, the search test 10× with 3 workers: 10 passed (54.4s) (third screenshot).
• ESLint (0 errors; 4 warnings in `CustomPropertiesPanel.spec.ts` that were already there) and Prettier are clean on the changed files.
• Before any fix, the same 10× run also passed 30/30, so the CI failures did not reproduce locally. The operator mis-pick did reproduce once, and its trace is what pinned down the asset-count race.

npx playwright test \
  playwright/e2e/Features/AIMode/CustomPropertiesPanel.spec.ts:305 \
  playwright/e2e/Pages/AIProfile/ProfileTeam.spec.ts:144 \
  playwright/e2e/Features/AIMode/PersonaSettings.spec.ts:939 \
  --project=chromium --repeat-each=10 --workers=3 --retries=0 --no-deps

Manual testing performed

None beyond the Playwright runs above.

UI screen recording:

Not applicable — test-only change.

Checklist:

• I have read the CONTRIBUTING document.
• I have commented on my code, particularly in hard-to-understand areas.
• I have added tests (unit / integration / Playwright as applicable) and listed them above.

🤖 Generated with Claude Code

[github-actions]

Hi there 👋 Thanks for your contribution!

The OpenMetadata team will review the PR shortly! Once it has been labeled as safe to test, the CI workflows
will start executing and we'll be able to make sure everything is working as expected.

Let us know if you need any help!

[github-actions]

🚦 Removed from the merge queue — failed_checks (2026-10-31T00:16:00Z)

Blocked the queue: playwright-summary

• Prune dead merge-queue refs — cancel-stale-runs
• Postgresql PR Playwright E2E Tests — playwright-summary, playwright / playwright-ci (chromium-09), playwright / playwright-ci (chromium-02), playwright / playwright-ci (chromium-06)

Failed Playwright tests:

• Features/OntologyFirstRun.spec.ts › creates an ontology from scratch and lands in Edit with the New class panel
• Features/OntologyNativeAuthoring.spec.ts › a competing lease leaves no partial class and the form stays usable
• Features/OntologyNativeAuthoring.spec.ts › creates every class field atomically, preserves free space, and cancels provisional relations
• Features/OntologyNativeExport.spec.ts › downloads every format of the current ontology release, then gates on SHACL
• Features/OntologyNativeStudio.spec.ts › authors a native model through the shared View, Edit, and Query shell

[gitar-bot]

Code Review ✅ Approved

🟢 Low risk · These changes adjust Playwright setup, waits, and test data isolation only, leaving application behavior and public interfaces unchanged.

Test-only de-flake of three Playwright specs: CustomPropertiesPanel edit now uses the isolated worksheet type, ProfileTeam search waits for each step's result before continuing, and PersonaSettings curated widget opens its deep link inside the running app and waits for the asset count before picking rule values. No issues found.

Verification

Rules No rules evaluated

Functional validation Not enabled

Set up

Auto-approval Not enabled

Set up

_{For more verification details, comment gitar display:verbose.}

Options

Display: compact → Showing compact verification results.

Comment with these commands to change the behavior for this request:

Compact

gitar display:verbose

_{Was this helpful? React with 👍 / 👎 | Powered by Gitar — free for open source}