Skip to content

What is Architecture Tracing?

Last updated

Architecture tracing is the analysis of how code changes propagate through a system's dependency graph — identifying which modules, services, and data flows are affected by a diff, and assessing the architectural risk of the change based on coupling, boundaries, and historical incident patterns.

Why does architecture tracing matter for engineering teams?

A few changed lines in a shared utility module can break many downstream consumers. Reviewers usually read the diff, not the code that depends on it, and tracing that by hand is slow. Architecture tracing makes the coupling visible before merge.

How does Argus handle architecture tracing?

A background worker parses the default branch into a file and symbol graph (go/ast for Go, tree-sitter for other supported languages, with regex fallbacks for TypeScript/JavaScript, Python, Java, Rust, and C#). For each PR, Argus walks that graph up to two levels back from the changed files and puts up to 50 dependents, plus the source of up to 3 direct dependents, into the reviewer model's prompt. A changed file with 5 or more dependent files is flagged as a choke point. The walk stays inside one repo, and the result is prompt context, not a risk score or a report posted on the PR.

Run Argus on your own repositories

Open source under AGPL-3.0, with no paid tier and no feature gating.

Self-hosted only: Docker Compose or Fly.io, Postgres with pgvector, a GitHub App and a Clerk app you create, your model keys and an embeddings endpoint.