Skip to main content

​
Traceability Matrix

The matrix is the canonical UI for traceability. One row per requirement, one column per test, one cell per link. The colour of each row tells you the coverage state at a glance; clicking a cell drills into the test run that produced the result.

​
Layout

The grid is rendered with TanStack Table — virtualised for projects with thousands of requirements, sticky headers, keyboard-navigable, and resizable columns. [Screenshot needed: matrix UI showing heatmap colouring, filter chips, and a hovered cell with run details]

​
Heatmap colour key

The legend at the top of the matrix is clickable — click gap to filter the view to only gap-state rows.

​
Filters

Filters compose. Stacking several narrows the view down to exactly what you need. Each filter renders as a chip at the top of the matrix. Click the X on a chip to remove it.

​
Saved views

Once you’ve composed a useful filter set, Save view persists it under a name (e.g. “ASIL-D regressions”). Saved views:
  • Are scoped to the project but visible to all members.
  • Get a direct URL you can paste in a Slack thread or PR comment.
  • Show in the sidebar of every team member’s matrix view.
A few teams use saved views as standing dashboards — “Safety leads — open this every Monday morning” — and link the URL in their team docs.

​
Sorts

Default sort is by requirement id. You can sort by:
  • id (lexicographic).
  • priority (critical first).
  • asil_pl (ASIL-D first).
  • coverage_state (gap, then regression, then partial, then unconfirmed, then confirmed).
  • last_changed (recency of any field edit).
  • last_verified (recency of last test run touching the requirement).
Hold shift to compose sorts — “first by ASIL descending, then by last verified ascending” is a good triage view.

​
Drill-in

Click any cell to open the Run detail drawer:
  • The test source code at the SHA, with the @confirms line highlighted.
  • The JUnit row — assertion message, duration, retries.
  • Attached artifacts — MCAP bag, logs, screenshots.
  • The runner metadata — image tag, hardware, wall-clock.
  • The requirement detail with full text on the right.
Click a row header (requirement) to open the Requirement detail page — the full text, history of edits, list of confirming tests, child requirements if any, and an audit trail.

​
Tree vs flat mode

Switch with the toggle in the toolbar. Both modes show identical data; the choice is purely visual.

​
Graph view

For visual reviewers, the Graph view renders the requirement → test graph as a force-directed diagram. Nodes are coloured by coverage state; edges are coloured by source (@confirms, manual, LLM-inferred). Graph view is best for:
  • Spotting clusters of gap-state requirements.
  • Reviewing the decomposition tree pre-audit.
  • Showing the matrix to non-engineering stakeholders.
It’s not a primary work surface — filter and sort on the table; explore on the graph.

​
Exports

Three formats, all from Matrix → Export: CSV is the most common — most regulated teams ingest it into their internal reporting before submission. ReqIF is for round-tripping back into the upstream requirements tool.

​
Multi-repo behaviour

For a project linked to multiple repos, the matrix gains a Repository column showing which repo a test lives in. Aggregation rules:
  • A requirement can be confirmed by tests in any repo in the project.
  • The Repository column is a filter target — drill the matrix to one repo for repo-scoped review.
See Multi-repo projects for the full model.

​
Snapshot mode

The toolbar has a snapshot picker. Default is Live (current state of the default branch). Switching to a snapshot — e.g. v1.2.0 — shows the matrix as it existed at the moment the snapshot was cut. Cells are read-only in snapshot mode. This is the view auditors see.

​
Performance

The matrix is virtualised — rendering 10,000 requirements × 20,000 tests is a few hundred milliseconds. Saved views and filters apply server-side for projects above 1,000 requirements to keep the wire small. For projects above 100,000 requirements (large standards-derivation projects), open a ticket — we have an indexed query mode that needs to be flipped on per project.

​
Next

Coverage

The state machine in detail.

Gaps

Filter to gap-only and close them.

Change-impact

How the PR Check Run highlights affected requirements.

Multi-repo

Rollup across many repos in one project.