* chore: added merge_group for github merge queue
* chore: remove unnecessary merger group on team labeler
* fix: added gates for merge queue and pull request events
The UI module's yarn preinstall runs js-antlr which requires the antlr4
CLI. This matches the pattern used in docker-openmetadata-server.yml.
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
The build-once optimization (#26423) used `-DonlyBackend -pl !openmetadata-ui`
which produces a tar.gz without the compiled React app. The Docker container
starts but cannot serve the login page, causing auth.setup.ts to timeout
on all 6 shards (input[id="email"] never appears).
The fix removes the backend-only flags so the full distribution including
the UI is built once and shared across shards.
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Previously each of the 6 Playwright shards independently ran the full
Maven build (~10-15 min each), wasting ~50 min of compute per CI run.
Extract Maven build into a dedicated `build` job that runs once:
- Builds with `mvn -DskipTests -DonlyBackend clean package`
- Uploads openmetadata-*.tar.gz as a GitHub Actions artifact
- Moves label check here to gate all downstream work
Shard jobs now:
- Download the pre-built artifact
- Pass `-s true` to run_local_docker.sh to skip Maven
- No longer need Maven cache
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Reverts the 8-shard configuration back to 6 shards. The dedicated
stateful/graph/landing shards increased infrastructure failure surface
without reducing test flakiness. This restores the simpler 6-shard
layout where shards 3-6 split chromium tests evenly.
Keeps the consolidated PR comment summary from #26418.
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Replace per-shard PR comments with a single consolidated comment posted
by a new playwright-summary job that runs after all shards complete.
Shows per-shard breakdown table, aggregated failures, and flaky tests
in one place. Cleans up old per-shard comments from previous runs.
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
* Improve Playwright CI: reduce shard contention and fix STEP_SUMMARY size
- Split 6 shards into 8 with dedicated projects for stateful (serial),
graph (2 workers), and landing (2 workers) tests to reduce contention
- Replace inline base64 screenshots in GITHUB_STEP_SUMMARY with a
lightweight table + artifact download link to stay under 1024k limit
- Add expect timeout of 15s for CI reliability
- Exclude landing page tests from the general chromium project to avoid
duplicate runs
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* Exclude stateful/graph/landing tests from chromium project
Prevent these tests from running twice — once in the dedicated
project shard and again in the general chromium shards.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
* Add Playwright failure reporting to PR comments
When a shard fails, automatically post a PR comment listing failed tests
with screenshot/trace availability and artifact download links. Updates
existing comments on re-runs to avoid duplicates.
Also uploads test-results (screenshots, traces) as separate artifacts
for easier debugging.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* Improve Playwright failure diagnostics
- Add JSON reporter for machine-readable test results
- Capture traces on every failure (not just retries) for debugging
- Record video on failure to see the full interaction sequence
- PR comment now distinguishes genuine failures vs flaky tests
- Genuine failures show error messages inline with expandable details
- Flaky tests listed separately so they're clearly not blockers
- Write failure screenshots to GitHub Step Summary for quick viewing
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* Address review: clean up stale comments and paginate lookup
- Delete failure comments when shard passes on re-run
- Use paginated comment lookup to handle PRs with 100+ comments
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
* Upgrade Playwright from 1.51.1 to 1.54.2
- Updated @playwright/test package to version 1.54.2 in package.json
- Updated Playwright browser installation to 1.54.2 in all GitHub workflow files:
- mysql-nightly-e2e.yml
- playwright-mysql-e2e.yml
- playwright-postgresql-e2e.yml
- postgresql-nightly-e2e.yml
- Updated yarn.lock to reflect the new Playwright version and its dependencies
Using 1.54.2 instead of 1.56.x to avoid regression bug with "Internal error: step id not found"
that affects fixtures in Playwright 1.55+ (see microsoft/playwright#37147).
This upgrade brings stable Playwright features, improvements, and bug fixes.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
* fix: update Playwright tests and utility functions for improved query handling and response validation
* fix: update URL wait conditions in Data Insight tests for accurate navigation
* fix: refine search response handling in verifyDeletedEntityNotVisible function
* fix: correct Playwright worker configuration in workflow files
- Fixed missing backslash continuation in playwright-postgresql-e2e.yml causing --workers=50% to be ignored
- Added missing --workers=50% flag in else block of playwright-postgresql-e2e.yml
- Fixed missing backslash continuation in postgresql-nightly-e2e.yml
- Fixed missing backslash continuation in mysql-nightly-e2e.yml
These syntax errors caused Playwright tests to run with unlimited workers (default CPU cores × 2),
leading to resource exhaustion and GitHub Actions runner communication failures. With this fix,
tests will properly run with 50% workers (1 worker on 2-core runners) preventing memory/CPU starvation.
* fix: increase Playwright worker count to 75% in workflow files
* fix: update Playwright tests to use correct role selectors and encode query parameters
* fix: reduce Playwright worker count from 75% to 50% in CI workflows
* fixed failing test
* fix: remove worker count specification from Playwright test commands in CI workflows
* fix: update Playwright test commands to specify worker count and adjust ESLint ignore settings
* fix: reduce worker count in Playwright configurations to prevent resource exhaustion
* fix: change waitForLoadState from 'networkidle' to 'domcontentloaded' in redirectToHomePage function
* fix: increase worker count to 3 in Playwright config and add navigation timeout to prevent hangs
* fix: reduce worker count to 2 in Playwright config to optimize resource usage
* Upgrade Playwright from 1.51.1 to 1.56.1
- Updated @playwright/test package to version 1.56.1 in package.json
- Updated Playwright browser installation to 1.56.1 in all GitHub workflow files:
- mysql-nightly-e2e.yml
- playwright-mysql-e2e.yml
- playwright-postgresql-e2e.yml
- postgresql-nightly-e2e.yml
- Updated yarn.lock to reflect the new Playwright version and its dependencies
- Reduced workers from 4 to 2 in playwright.config.ts to prevent resource exhaustion
- Added navigationTimeout: 60000 to prevent infinite hangs on networkidle waits
- Changed networkidle to domcontentloaded in redirectToHomePage function
This upgrade brings the latest stable Playwright features, improvements, and bug fixes while addressing resource exhaustion issues.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix: downgrade Playwright to version 1.55.1 in workflow files and package.json
* chore: upgrade Playwright to version 1.57.0 in workflow files and package.json
* trying firfox
* chore: upgrade Playwright to version 1.57.0 in OpenMetadata
- Updated @playwright/test package to version 1.57.0 in package.json
- Updated Playwright browser installation to 1.57.0 in all GitHub workflow files:
- mysql-nightly-e2e.yml
- playwright-postgresql-e2e.yml
- postgresql-nightly-e2e.yml
- Updated yarn.lock to reflect the new Playwright version and its dependencies
- Maintained Chromium browser configuration for all test projects
- Kept worker optimization at 2 workers and --workers=50% flag for resource management
This upgrade brings the latest stable Playwright features, improvements, and bug fixes, including resolution of the "step id not found" regression bug.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* chore: update Playwright trace configuration to 'on-first-retry'
* revert worker setting
* fix failing test
---------
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Sriharsha Chintalapani <harshach@users.noreply.github.com>
* data asset rule enforcement on ui
* added multi user and multi team support at same time
* fix the multi user case which wasn't working correctly and clean the ui as well
* fix the rule condition for glossary
* fix the glossary dropdown getting closed on select select after selection where we should wait for footer dropdown button action
* added playwright test for data asset rules table
* fix unit test and supported playwright test for table entity
* change the method and variable name for better readability
* naming convention fix around the glossary validation rule
* supported playwright test for all the entities
* supported in service entities
* spearate the test of rules enable and disable state and added supported in glossaryand teams page of rules owner
* Update generated TypeScript types
* fix the unit test and supported rule in tags page
* fix the generated type issue
* fix he domain single select where de-selecting throwing error and fix type
* modify the playwrigth config so dataAssetRules should be run after the chromium is finish executing so the global run changes won't interrupt other test
* fix all unit test
* fix the test failing in dataAssetHeader
* trying the chromium parallel test running
* modifyed yaml files to run the data asset rules after all machine runs is completed
* change the test running order in playwright config
* reverted the playwright config
* run the data asset rules in only one runner
* added support for multiple user and team combination in ownerLabel compoennt
* fix the placemnet issue, reduce the team font to 12 and minor cleanup
* modify many owner component cases as per ui
* clear ruleEnforcementProvider and useEntityRules hooks
* change the dropdown on owner user to mui component
* fix unit test and make chnages in playwright postgress around test running
* fix unit test failing
* fix some playwright test due to rule enforcement and made changes in CI playeright config to only run dataasset on container 6
* reverted the baseurl
* fix the version view for owner, fix some failing playwright test and re-modify the postgress script to run test
* supported the placement option for owner in case of user and team both and postgress e2e change
* remove the other e2e file changes and fix playwright around rules changes
* fix the e2e command
* fix the failing test around disabled rule and based on comments made fix
---------
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
* Upgrade playwright version to 1.54.2 (latest) to support indexedDb storageState (#22879)
* bump playwright version to latest
* update playwright version in workflow files
* update playwright installation step in workflow files to ensure consistent working directory
---------
Co-authored-by: “maulikkotak98” <“maulik.k@deuexsolutions.com”>
* update playwright version to 1.51.1
* fix tests for lineage
* minor refactor and cleanup
* fix lineage tests
---------
Co-authored-by: Maulik Kotak <72746599+maulikkotak98@users.noreply.github.com>
Co-authored-by: “maulikkotak98” <“maulik.k@deuexsolutions.com”>
Co-authored-by: Aniket Katkar <aniketkatkar97@gmail.com>
* Add AutoPilot playwright tests
* Update the playwright config and separate the nightly tests
* Delete the created service
* Fix the tests
* Increase the AutoPilot timeout
* Reduce the timeout for AutoPilot and add default filter inputs
* - Update playwright workflow
- Move the ingestion based tests unaffected by the autopilot app out of nightly
* Renamed nightly folder to follow the coding standard
---------
Co-authored-by: Shailesh Parmar <shailesh.parmar.webdev@gmail.com>
* [CHORE] Move workflows to larger runners
* chore: Move only playwright tests to large runners
---------
Co-authored-by: Akash-Jain <15995028+akash-jain-10@users.noreply.github.com>
* fix(e2e): ingestion related flaky failures
* no need to run workflow for ready_for_review state update
* await for queued or running status
* fix failing tests
* fix issue
* add support for partial success
