docs(flaky-tests): document failing tests list in failure count monitor preview panel

[samgutentag]

Summary

Documents the searchable Failing tests list added to the Failure Count monitor preview panel. The list shows per-test failure counts, highlights counts that meet the configured minimum threshold in red, supports case-insensitive search by test name or parent name, and links each test name to its detail page.

Source

• trunk2 PR: https://github.com/trunk-io/trunk2/pull/3620

Test plan

• Preview in GitBook

[trunk-staging-io]

😎 Merged successfully - details.

[claude]

Claude finished @samgutentag's task in 46s —— View job

---

Review complete

• Read CLAUDE.md for repo-specific guidelines
• Read changed files
• Review for typos, grammar, and formatting issues
• Review for documentation best practices
• Post inline comments and final verdict

---

The new ## Preview Panel section is well-written, clear, and consistent with the rest of the document's style and voice. One minor suggestion posted inline:

failure-count-monitor.md line 101 — "minimum failure threshold" is a one-off term; the document consistently uses "failure count" (the configuration section header and body both use that phrase). Swapping in "configured failure count" keeps terminology aligned.

Everything else looks good: no typos or grammar issues, proper use of bold for UI elements, correct placement within the document flow, and the 100-result truncation behavior and "No tests match" empty-state are both well-documented.

[github-actions]

Auto-approved: Claude code review passed.

[samgutentag]

Triage status (2026-05-05): up to date with main. Needs review before merge.

[claude]

Claude finished @samgutentag's task in 1m 9s —— View job

---

Review complete

• Read CLAUDE.md for repo-specific guidelines
• Read changed files and PR diff
• Review for typos, grammar, and formatting issues
• Review for documentation best practices
• Check whether prior review feedback was addressed
• Post final verdict

---

The new ## Preview Panel section is well-written and consistent with the rest of the document's voice and structure. Section heading uses Title Case to match neighbors, bold is applied correctly to UI strings (Preview, Failing tests), and the 100-result truncation and "No tests match" empty-state behaviors are clearly documented. The phrase "detail page" matches usage in test-labels.md, test-collections.md, and managing-detected-flaky-tests.md.

Outstanding from the prior review (non-blocking)

The terminology nit posted on the previous review is still unresolved on flaky-tests/detection/failure-count-monitor.md:101: the new section uses "minimum failure threshold," but the rest of the document — including the ## Failure Count configuration section header and body — consistently uses "failure count." Since the recent commits were just a merge from main and a triage status comment, the prior suggestion didn't get applied.

Suggested edit to align terminology:

Counts that meet or exceed your configured failure count are highlighted in red; counts below the threshold appear in muted text.

Fix this →

This is a minor consistency suggestion, not a blocking issue — the PR is otherwise good to merge.
· Branch

[github-actions]

Auto-approved: Claude code review passed.

[samgutentag]

Verification status (2026-05-07, re-verified): live

Verified: customers can use this. Ready to publish.

• Eng PR: trunk2#3620 ("feat: add failing tests list to failure count preview panel", merged 2026-04-21, merge commit still on main)
• Linear ticket: none referenced
• Flag: none (no LaunchDarkly gate found in PR body or diff; UI shipped unflagged)
• Signals: pure frontend addition, no rollout to chase.

Review nit applied in commit b64dc6e:

• Line 101 wording aligned with the rest of the page. "minimum failure threshold" replaced with "configured failure count" so the preview-panel section uses the same terminology as the ## Failure Count section header and body. Also dropped the second occurrence of "threshold" in the same sentence (now "counts below appear in muted text") since the configured failure count is the only number being compared against.

trunk check clean on the edited file.

Suggested next action: flip out of draft and merge.

[claude]

Review in progress

• Read CLAUDE.md for repo-specific guidelines
• Read changed files and PR diff
• Review for typos, grammar, and formatting issues
• Review for documentation best practices
• Check whether prior review feedback was addressed
• Post final verdict

View job run

[claude]

Claude finished @samgutentag's task in 1m 19s —— View job

---

Review complete

• Read CLAUDE.md for repo-specific guidelines (no CLAUDE.md found in repo)
• Read changed files and PR diff
• Review for typos, grammar, and formatting issues
• Review for documentation best practices
• Check whether prior review feedback was addressed
• Post final verdict

---

The new ## Preview Panel section is well-written, accurate, and consistent with the rest of the document.

Prior review feedback resolved. Commit b64dc6e aligned line 101's terminology with the rest of the page: "minimum failure threshold" → "configured failure count," and the trailing "below the threshold" was simplified to "below appear in muted text." Both swaps match the established ## Failure Count section heading and body, so the page now reads consistently end-to-end.

Style and formatting checks

• Section heading: ## Preview Panel uses Title Case to match neighbors (## When to Use This Monitor, ## Detection Type, ## How It Works, ## Resolution Behavior, ## Choosing Between Monitors).
• Bold for UI strings: Preview and Failing tests are correctly emphasized; matches the convention used elsewhere in the document for UI labels (e.g., flaky, broken).
• Terminology: "detail page" matches usage in test-labels.md, test-collections.md, and managing-detected-flaky-tests.md.
• Placement: positioned between ## Resolution Behavior and ## Muting, which fits the document's flow (configuration/behavior topics → UI affordances → meta-actions).
• Typos/grammar: none found.
• Coverage: 100-result truncation, "No tests match" empty-state, case-insensitive search, and parent-test-name search are all documented.

No blocking issues. PR is ready to merge.
· Branch

[github-actions]

Auto-approved: Claude code review passed.

[claude]

Claude finished @samgutentag's task in 2m 15s —— View job

---

Review complete

• Read CLAUDE.md for repo-specific guidelines (no CLAUDE.md found)
• Read changed files and PR diff
• Review for typos, grammar, and formatting issues
• Review for documentation best practices
• Check whether prior review feedback was addressed
• Post final verdict

---

Blocking: duplicate ## Preview Panel H2 section

The merge from main in commit 60273c4 pulled in PR #555, which also added a ## Preview Panel section to flaky-tests/detection/failure-count-monitor.md (with ### Status Filter and ### Large Repo Truncation subsections). The merge did not unify the two — the file now has two H2 headings with the same name:

$ grep -n "^## " flaky-tests/detection/failure-count-monitor.md
...
 89:## Resolution Behavior
 97:## Preview Panel        ← this PR's new section
105:## Muting
109:## Preview Panel        ← from #555, came in with the merge
113:### Status Filter
125:### Large Repo Truncation
129:## Choosing Between Monitors

This is a regression introduced by the merge, not by the original PR commits — earlier Claude reviews ran before #555 landed on main, so they couldn't have caught it. But it must be resolved before merging: two identically-named H2s break the table of contents / anchor links and confuse readers.

Suggested resolution

Merge the two sections into a single ## Preview Panel block. A natural shape:

## Preview Panel

When you create or edit a failure count monitor, a **Preview** panel appears on the
right side of the dialog on larger screens. The preview updates as you adjust the
monitor's settings, giving you a live look at what the monitor would detect against
your current branch data.

### Failing Tests List

Once the monitor configuration produces detections, the panel shows a **Failing tests**
list. Each row displays the test name as a link to its detail page, along with its
failure count. Counts that meet or exceed your configured failure count are highlighted
in red; counts below appear in muted text.

You can search the list by test name or parent test name. The search is
case-insensitive and filters as you type. If no tests match your search term, the list
shows a "No tests match" message. When more than 100 tests are detected, only the
first 100 are shown with a notice to narrow your search.

### Status Filter

A **status filter dropdown** in the preview panel lets you filter the test list to any
combination of statuses: **Healthy**, **Flaky**, and **Broken**. By default, all
statuses are shown.

[…rest of existing Status Filter content unchanged…]

### Large Repo Truncation

[…existing content unchanged…]

Note that #555's intro sentence ("a preview panel shows which tests the current configuration would flag based on recent data") is now redundant with this PR's intro paragraph, so it can be dropped — the new wording is more specific and complete. The ### Large Repo Truncation subsection is also somewhat redundant with this PR's new "more than 100 tests" sentence; consider whether both are needed or if the truncation behavior should be folded into the Failing Tests List paragraph instead.

Fix this →

Non-blocking observations

The PR's new content itself is well-written:

• Heading style — ## Preview Panel uses Title Case to match the rest of the page.
• UI emphasis — Preview and Failing tests are correctly bolded; matches the convention used for flaky / broken elsewhere.
• Terminology — "configured failure count" (line 101) was correctly aligned with the ## Failure Count configuration section in commit b64dc6e. "detail page" matches usage in test-labels.md, test-collections.md, and managing-detected-flaky-tests.md.
• Coverage — 100-result truncation, "No tests match" empty-state, case-insensitive search, and parent-test-name search are all clearly documented.
• Typos/grammar — none found in the new content.

Verdict

Not ready to merge. The duplicate ## Preview Panel H2 must be resolved (the new content needs to be folded into the existing section that came in via main). Once that's done, the page is good to go.
· Branch