Backport Mlflow AutoPilot E2E fix to 1.13

[IceS2]

Backports the Playwright half of #30191 to 1.13. 1.13 carries the same mlflow-skinny>=3.13.0,<3.15 bump, so the AutoPilot E2E fails here the same way.

What

MlFlowIngestionClass filled placeholder connection strings that MLflow resolved to a local file store, so the agent walked an empty registry and the test passed. MLflow 3.13 put the file store into maintenance mode and now raises instead:

MlflowException: The filesystem tracking backend (e.g., './mlruns') is in maintenance mode
and will not receive further updates. ... set `MLFLOW_ALLOW_FILE_STORE=true` to opt out.

The agent throws on connect, the AutoPilot workflow never completes successfully, and the run-completed banner never renders. Verified against the pinned version:

mlflow-skinny	search_registered_models() with a placeholder URI
3.11.1 (old ceiling)	OK — file store accepted
3.13.0 (new floor)	raises MlflowException

The fields now hold sqlite DSNs. SQLite is a supported backend that bootstraps its own schema, so it keeps the "empty registry, agent succeeds" semantics the test already relied on — as a deliberate choice rather than a side effect of MLflow's string parsing. Both fields are plain type: string in the schema, so no schema or UI change is needed.

Scope

Playwright only, and intentionally byte-identical to the change already merged on main so the branches do not diverge.

Greptile flagged that the fixed /tmp paths are shared across services in the ingestion container (possible lock contention on concurrent first-time schema creation). That is a fair point but applies equally to main, which already merged these paths — so it should be hardened there first and backported, not introduced here as a divergence.

The connector-side version-selection fix from #30191 and its regression test are not included here, by request — 1.13 still carries that bug (a registered model is dropped after its description/tags/aliases are edited). The test change is the regression guard for that fix, so both were dropped together; keeping the test without the code fix would leave a failing integration test.

Greptile Summary

This PR updates the MLflow Playwright setup for MLflow 3.13 compatibility. The main changes are:

• Replaces placeholder tracking and registry URIs with SQLite DSNs.
• Preserves the empty-registry behavior without requiring a live MLflow service.
• Limits the backport to the Playwright fixture.

Confidence Score: 5/5

No additional blocking issue qualifies for this follow-up review.

• No distinct unresolved failure remains after excluding the already reported SQLite isolation issue.

Important Files Changed

Filename	Overview
openmetadata-ui/src/main/resources/ui/playwright/support/entity/ingestion/MlFlowIngestionClass.ts	Updates the MLflow Playwright connection fields to use SQLite tracking and registry backends supported by MLflow 3.13.

_{Reviews (3): Last reviewed commit: "fix(ui): unbreak the Mlflow AutoPilot E2..." | Re-trigger Greptile}

Context used:

• Context used - CLAUDE.md (source)
• Context used - openmetadata-ui-core-components/CLAUDE.md (source)

[github-actions]

❌ PR checklist incomplete

This PR cannot be merged until the following are addressed on its linked issue:

• No GitHub issue is linked. Link an issue in the Development section of the PR (or add Fixes #12345 to the description). For a same-org cross-repo issue, add Fixes open-metadata/<repo>#123 to the description.

The fields live on the linked issue in the Shipping project (open the issue → right sidebar → Projects). After you set them, re-run this check (or push a commit) — issue/project changes do not re-trigger it automatically.

Maintainers can bypass this check by adding the skip-pr-checks label.

[github-actions]

The Python checkstyle failed.

Please run make py_format and py_format_check in the root of your repository and commit the changes to this PR.
You can also use pre-commit to automate the Python code formatting.

You can install the pre-commit hooks with make install_test precommit_install.

[sonarqubecloud]

Quality Gate passed for 'open-metadata-ingestion'

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
100.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[github-actions]

🟡 Playwright Results — all passed (11 flaky)

✅ 3968 passed · ❌ 0 failed · 🟡 11 flaky · ⏭️ 30 skipped

Shard	Passed	Failed	Flaky	Skipped
✅ Shard 1	291	0	0	4
✅ Shard 2	0	0	0	0
🟡 Shard 3	735	0	2	7
🟡 Shard 4	780	0	3	2
🟡 Shard 5	754	0	3	9
✅ Shard 6	693	0	0	0
🟡 Shard 7	715	0	3	8

🟡 11 flaky test(s) (passed on retry)

• Features/BulkEditEntity.spec.ts › Database Schema (shard 3, 1 retry)
• Features/BulkImport.spec.ts › Database service (shard 3, 1 retry)
• Features/Permissions/GlossaryPermissions.spec.ts › Team-based permissions work correctly (shard 4, 1 retry)
• Flow/ExploreAggregationCountsMatching.spec.ts › should verify left panel counts and tab search results for normal search (shard 4, 2 retries)
• Flow/PersonaFlow.spec.ts › Set default persona for team should work properly (shard 4, 1 retry)
• Pages/CustomProperties.spec.ts › Enum (shard 5, 1 retry)
• Pages/DataContracts.spec.ts › Create Data Contract and validate for Spreadsheet (shard 5, 1 retry)
• Pages/Domains.spec.ts › Rename domain with subdomains attached verifies subdomain accessibility (shard 5, 1 retry)
• Pages/Lineage/LineageFilters.spec.ts › Verify lineage schema filter selection (shard 7, 1 retry)
• Pages/TestSuite.spec.ts › Logical TestSuite (shard 7, 1 retry)
• Pages/Users.spec.ts › Permissions for table details page for Data Consumer (shard 7, 1 retry)

📦 Download artifacts

How to debug locally

# Download playwright-test-results-<shard> artifact and unzip
npx playwright show-trace path/to/trace.zip    # view trace

[gitar-bot]

⚠️ Bug: SQLite DB paths are shared, not scoped per service instance

The PR description states the SQLite paths are "scoped per service instance so concurrent runs sharing the ingestion container cannot contend on one file or inherit a half-created schema from an earlier attempt," but the committed code uses fixed paths for the tracking and registry databases. Any parallel Playwright workers/services or a retry after a partial failure will contend on the same files or inherit a half-created schema, reintroducing exactly the flakiness the change claims to prevent. Since this.serviceName (which embeds a uuid()) is available on the base class, derive the DB filenames from it, e.g., sqlite:////mlflow_tracking_${this.serviceName}.db.

Scope both SQLite backends per service instance using the unique serviceName.:

await page.fill(
  '#root\/trackingUri',
  `sqlite:////tmp/mlflow_tracking_${this.serviceName}.db`
);
await checkServiceFieldSectionHighlighting(page, 'trackingUri');
await page.fill(
  '#root\/registryUri',
  `sqlite:////tmp/mlflow_registry_${this.serviceName}.db`
);
await checkServiceFieldSectionHighlighting(page, 'registryUri');

_{Was this helpful? React with 👍 / 👎}

[github-actions]

The Python checkstyle failed.

Please run make py_format and py_format_check in the root of your repository and commit the changes to this PR.
You can also use pre-commit to automate the Python code formatting.

You can install the pre-commit hooks with make install_test precommit_install.

[gitar-bot]

Code Review ⚠️ Changes requested 0 resolved / 1 findings

Backports MLflow AutoPilot E2E test fixes to 1.13, but the implementation uses fixed paths instead of service-scoped SQLite database paths, risking collision between concurrent test runs.

⚠️ Bug: SQLite DB paths are shared, not scoped per service instance

📄 openmetadata-ui/src/main/resources/ui/playwright/support/entity/ingestion/MlFlowIngestionClass.ts:61 📄 openmetadata-ui/src/main/resources/ui/playwright/support/entity/ingestion/MlFlowIngestionClass.ts:63

The PR description states the SQLite paths are "scoped per service instance so concurrent runs sharing the ingestion container cannot contend on one file or inherit a half-created schema from an earlier attempt," but the committed code uses fixed paths for the tracking and registry databases. Any parallel Playwright workers/services or a retry after a partial failure will contend on the same files or inherit a half-created schema, reintroducing exactly the flakiness the change claims to prevent. Since this.serviceName (which embeds a uuid()) is available on the base class, derive the DB filenames from it, e.g., sqlite:////mlflow_tracking_${this.serviceName}.db.

Scope both SQLite backends per service instance using the unique serviceName.

await page.fill(
  '#root\/trackingUri',
  `sqlite:////tmp/mlflow_tracking_${this.serviceName}.db`
);
await checkServiceFieldSectionHighlighting(page, 'trackingUri');
await page.fill(
  '#root\/registryUri',
  `sqlite:////tmp/mlflow_registry_${this.serviceName}.db`
);
await checkServiceFieldSectionHighlighting(page, 'registryUri');

🤖 Prompt for agents

Code Review: Backports MLflow AutoPilot E2E test fixes to 1.13, but the implementation uses fixed paths instead of service-scoped SQLite database paths, risking collision between concurrent test runs.

1. ⚠️ Bug: SQLite DB paths are shared, not scoped per service instance
   Files: openmetadata-ui/src/main/resources/ui/playwright/support/entity/ingestion/MlFlowIngestionClass.ts:61, openmetadata-ui/src/main/resources/ui/playwright/support/entity/ingestion/MlFlowIngestionClass.ts:63

   The PR description states the SQLite paths are "scoped per service instance so concurrent runs sharing the ingestion container cannot contend on one file or inherit a half-created schema from an earlier attempt," but the committed code uses fixed paths for the tracking and registry databases. Any parallel Playwright workers/services or a retry after a partial failure will contend on the same files or inherit a half-created schema, reintroducing exactly the flakiness the change claims to prevent. Since `this.serviceName` (which embeds a `uuid()`) is available on the base class, derive the DB filenames from it, e.g., `sqlite:////mlflow_tracking_${this.serviceName}.db`.

   Fix (Scope both SQLite backends per service instance using the unique serviceName.):
   await page.fill(
     '#root\/trackingUri',
     `sqlite:////tmp/mlflow_tracking_${this.serviceName}.db`
   );
   await checkServiceFieldSectionHighlighting(page, 'trackingUri');
   await page.fill(
     '#root\/registryUri',
     `sqlite:////tmp/mlflow_registry_${this.serviceName}.db`
   );
   await checkServiceFieldSectionHighlighting(page, 'registryUri');

Options

Display: compact → Showing less information.

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

Compact

gitar display:verbose

_{Was this helpful? React with 👍 / 👎 | Gitar}