test: expand E2E coverage with 30 new Playwright tests

[teeohhem]

Summary

Expands Playwright E2E test coverage from 88 to 118 tests (+30 net new), covering core user workflows across alerts, chart explorer, dashboards, search, services, sessions, sources, traces, and user preferences.

What changed

New test coverage (30 unique tests across 10 feature areas):

• Alerts: Alert card state badge, multi-alert ordering, shared createSavedSearchAlert helper
• Chart explorer: Chart type tabs, group-by series, metric source queries, SQL mode
• Dashboards: Section creation, tile editing, dashboard deletion from listing page
• Dashboard list: Search dashboards by name (new feature page)
• Search: Query mode preservation (SQL↔Lucene round-trip), empty results handling, Escape to close side panel
• Services dashboard: Top endpoints table, endpoint click navigation
• Sessions: Multiple session cards, session details drawer, data source filtering
• Sources: Source listing, source type switching, edit sources navigation
• Traces: Trace search/filter, side panel tabs, waterfall view with trace attributes
• Temporary dashboard: Listing page navigation, URL param persistence
• User preferences: Menu navigation, preferences modal, team settings, Escape to close

Consolidation (7 files deleted, no coverage lost):

• Merged 6 pairs of split spec files (*-extended, *-workflow, *-management) into single files per feature area
• Removed 12 tests that duplicated behavior already covered by existing comprehensive tests

Infrastructure:

• Added data-testid attributes to AlertsPage, DBEditTimeChartForm, SearchWhereInput
• Added page object methods: ChartExplorerPage.getChartTypeTab(), ServicesDashboardPage.topEndpointsTable, SidePanelComponent.getTabPanel()
• Created UserPreferencesPage and AlertsPage page objects
• Updated CI to fail E2E report when a shard is cancelled

Test plan

• Run full E2E suite: ./scripts/test-e2e.sh
• Verify no test name collisions across spec files
• Verify CI sharding distributes new tests evenly

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
hyperdx-oss	Ready	Preview, Comment	Mar 26, 2026 9:10pm

[changeset-bot]

⚠️ No Changeset found

Latest commit: c2575f2

Merging this PR will not cause a version bump for any packages. If these changes should not result in a new version, you're good to go. If these changes should result in a version bump, you need to add a changeset.

This PR includes no changesets

When changesets are added to this PR, you'll see the packages that this PR includes changesets for and the associated semver types

Click here to learn what changesets are, and how to add one.

Click here if you're a maintainer who wants to add a changeset to this PR

[github-actions]

PR Review

• ⚠️ createSavedSearchAlert helper unused in third test (alerts.spec.ts:336) → The "should display multiple alerts and verify ordering" test duplicates ~40 lines of alert-creation setup inline instead of calling createSavedSearchAlert() twice. Use the helper to stay DRY.

• ⚠️ Misleading test name (alerts.spec.ts:336) → Test claims to "verify ordering" but only asserts both alerts are visible (count() >= 2). Either verify actual ordering or rename to "should display multiple alerts".

• ⚠️ Fragile locators in traces spec (traces.spec.ts) → locator('[role="button"]').filter({ hasText: /\w+/ }) and locator('div[class*="HyperJson_key__"]') are overly broad/CSS-module-dependent. These existed before but are being promoted into the consolidated spec — worth replacing with data-testid attributes.

✅ No critical bugs or security issues. The data-testid attributes added to source all match what the new page objects/tests query. CI change for cancelled shard handling is correct.

[github-actions]

Knip - Unused Code Analysis

⚪ No changes detected (182 issues on both main and PR)

What is this?

Knip finds unused files, dependencies, and exports in your codebase.
This comment compares the PR branch against main to detect regressions.

Run yarn knip locally to see full details.

[github-actions]

E2E Test Results

✅ All tests passed • 125 passed • 3 skipped • 1090s

Status	Count
✅ Passed	125
❌ Failed	0
⚠️ Flaky	3
⏭️ Skipped	3

Tests ran across 4 shards in parallel.

View full report →

[teeohhem]

@claude are you able to use your new pull request review feature on this PR?

[github-actions]

Claude finished @teeohhem's task in 2m 12s —— View job

---

PR Review

• Gather context
• Read changed files
• Review test quality and patterns
• Check for fragile locators and anti-patterns
• Verify page object usage and structure
• Post comprehensive review

---

Overall this is solid work — the consolidation is good, the new page objects are clean, and all 125 tests passed in CI. A few issues worth addressing before merge:

Issues

1. createSavedSearchAlert helper unused in "verify ordering" test (alerts.spec.ts:162–228)

The bot already flagged this. The "should display multiple alerts and verify ordering" test duplicates ~40 lines of alert-creation setup inline instead of calling createSavedSearchAlert() twice. The helper exists precisely for this. Fix this →

2. Misleading test name (alerts.spec.ts:162)

"should display multiple alerts and verify ordering" but expect(await alertCards.count()).toBeGreaterThanOrEqual(2) only asserts both alerts are visible — it doesn't verify ordering. Rename to "should display multiple alerts" or add an ordering assertion (e.g., verify alert A appears before alert B in the DOM). Fix this →

3. Fragile locators in traces waterfall test (traces.spec.ts:97–123)

Two locators that will be brittle:

// Line 98 — too broad: matches any button with any text anywhere on the page
locator('[role="button"]').filter({ hasText: /\w+/ })

// Line 121 — depends on CSS module hash in class name, breaks if build config changes
locator('div[class*="HyperJson_key__"]').filter({ hasText: new RegExp(`^${attribute}$`) })

For the trace timeline buttons, add a data-testid="trace-waterfall-span" and use getByTestId(). For HyperJson_key__, add a data-testid="hyperjson-key" to the HyperJson component. Fix this →

4. Single-test files in search/ (search/search-modes.spec.ts, search/search-table.spec.ts, search/side-panel.spec.ts)

Each of these new files contains exactly 1 test. The added overhead (3 describe blocks, 3 beforeEach setups, 3 file paths to maintain) isn't worth isolating 3 tests. These belong in the existing search/search.spec.ts (for side panel and modes) and search/search-filters.spec.ts (for empty results), or at minimum collapsed into a single new search/search-ui.spec.ts. Fix this →

5. Missing waitForRowsToPopulate() before clickFirstRow() (search/side-panel.spec.ts:14)

await searchPage.table.clickFirstRow();  // no wait — first row may not exist yet

Other specs that open the side panel (e.g. search.spec.ts) call submitEmptySearch() first to ensure data is loaded. This test skips directly to clicking, which will be flaky on a cold start. Fix this →

6. Page object bypassed for data-testid lookup (search/search-table.spec.ts:20)

const noResults = searchPage.page.getByTestId('db-row-table-no-results');

Direct getByTestId in a spec file bypasses the page object layer. This belongs in SearchPage or its TableComponent. Fix this →

Minor

• Stale AlertsPage.ts comment (AlertsPage.ts:4): "Currently not used until Alerts tests are implemented" — alerts tests are now implemented, remove the comment.

• sessions.spec.ts:37 — locator('role=dialog') should be getByRole('dialog'). The locator() form with role selector is legacy Playwright syntax.

• sessions.spec.ts:57–68 ("should filter sessions by selecting data source") is nearly identical to the test at line 15–25 ("should display multiple session cards") — both select the source and assert a card is visible. The distinction is thin enough to consider merging.

What's good

• The consolidation of *-extended/*-workflow/*-management split files is clean and removes duplication without losing coverage.
• UserPreferencesPage and AlertsPage page objects are well-structured with proper encapsulation.
• data-testid additions to AlertsPage.tsx (alert-card-{id}, alert-link-{id}, alert-state-badge) are well-placed and scoped correctly.
• SidePanelComponent.getTabPanel() is a clean addition.
• CI shard cancellation fix in main.yml is correct.
• The SQL query in chart-explorer.spec.ts tests a realistic workflow with proper parametrized timestamps.