Skip to Content
FeaturesCode Health Dashboard

Code Health Dashboard

The Code Health page at Dashboard → Code Health is structured around three layers, top to bottom. Most users get their answer from the top one.

Layer 1 — Verdict (top of page)

A single sentence answers “should I be worried?”:

  • All clear — quality gate passing, score steady or improving
  • Code health is degrading — score trending down (review recent commits)
  • Quality gate failed — one or more conditions not met, listed inline

The verdict is color-coded and includes a “View failing checks” button that jumps directly to the Issues tab when conditions fail. No clicks needed to see why the gate failed; the conditions are inline-expandable.

This replaces the older small “Quality Gate” pill in the header that required a click to reveal conditions. If you’re checking status quickly, the verdict is the only thing you need to read.

Layer 2 — What changed since previous scan

A delta block immediately below the verdict shows:

  • Score change — e.g. 78 → 72 (-6) since the previous scan, with a trend arrow
  • New issues — count of issues introduced since the previous scan
  • Trend direction — improving / stable / degrading

Scope toggle: All issues vs New code only

Right side of the delta block has a two-button toggle:

  • All issues (default) — every finding in the latest scan, regardless of when it was introduced
  • New code only — only issues introduced since the previous scan

When you click “New code only”, the entire dashboard re-scopes — metric counts, hotspots, top issues all reflect just what changed. This answers the engineer’s daily question “did I just break something?” without requiring you to compare scans manually.

Note: The toggle is disabled if no comparable previous scan exists yet. Run at least two scans to enable new-code mode.

Layer 3 — Hero metric + supporting tiles

Below the delta block, five metric tiles:

  1. Quality Score (hero card on left) — composite 0-100 score, with delta vs previous scan
  2. Security & Reliability (clickable, opens repo dashboard) — combined ratings
  3. Maintainability (clickable, opens code smells page) — rating + smell count + MI Index
  4. Risk Density — severity-weighted issue concentration
  5. Duplication & Dead Code — combined into one tile

The previous design had 8 tiles in a rainbow strip. The new layout puts visual weight on the score (the headline number) and groups related metrics into combined tiles.

Layer 4 — Drill-down tabs

Three tabs, each focused on one job:

Hotspots tab (default)

The riskiest files, sorted by severity-weighted issue count:

  • File path, complexity score, issue count, last-modified date
  • AI actions per row: Fix with AI, Document, Generate Tests
  • Output panels render below the file row

Top Issues tab

Severity-prioritized list of up to 10 issues across bugs, vulnerabilities, and code smells:

  • Color-coded severity badges
  • File:line reference
  • Inline message
  • “Quick fix” / “Requires review” effort estimate
  • One-click Fix with AI per issue
  • Technical Debt Trend chart — quality score and issue count over time
  • Smell Categories breakdown — code smell distribution by category

AI fix caching

Every “Fix with AI” click on this page is cached per finding. Subsequent clicks on the same finding return instantly without consuming AI quota. The toast message tells you which path was taken:

  • “Cached fix loaded — no AI quota used” — instant return, free
  • “AI fix generated” — first generation, stored for future clicks
  • “Fresh fix generated” — explicit regenerate

See AI Fixes & Caching for the full caching reference.

Quality gate

The gate verdict in the banner reflects the same gate configured at Settings → Policies → Quality Gates. Configure thresholds there; this page surfaces the result.

Failed gate conditions are listed inline in the verdict banner. Click a condition to navigate to the relevant view (e.g. failing security rating opens the security findings list).

Repository and branch

  • Repository selector — top-right of the header. Switching repos re-fetches all data.
  • Branch selector — when supported, shows quality metrics for a specific branch. Defaults to the repository’s default branch.

Export

Top-right export button downloads a JSON snapshot of every metric on the page including the delta vs previous scan, hotspot list, top issues, and trend series. Use for compliance reports or external dashboards.

Permissions

ActionRequired Role
View dashboardAny role
Trigger AI fix / document / generate testsMember or higher (counts against AI remediation quota)
Configure quality gate thresholdsOrg Admin or Owner (handled at Settings → Policies → Quality Gates)

Behavior notes

  • Loading state uses skeleton placeholders matching the layout — no centered spinner.
  • Tab state is local to the page; navigating away resets to the Hotspots tab on return.
  • Scope state (overall vs new) persists across repo switches in the same session.
PR ReviewsPR Reviews — Detailed Guide