n8n/packages/testing
Declan Carroll 8b93863eee
fix(coverage): Exclude test files, mocks, and .d.ts from coverage collection (no-changelog)
Four sources of noise in coverage scores:
- Jest root config: src/**/*.ts captured test files co-located in __tests__/
  and declaration files (.d.ts) alongside production code
- nodes-base: credentials/**/  nodes/**/ patterns captured 853 .spec.ts files,
  dragging function coverage scores with 0%-covered test code
- Monocart sourceFilter: no .d.ts or test file exclusion after source-map
  expansion in E2E coverage
- vitest-config/frontend CI: src/** (broad) narrowed to src/**/*.{ts,vue}
  to match the local non-CI default

Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>
2026-06-08 20:30:39 +01:00
..
code-health refactor: Consolidate janitor + code-health onto a shared rules-engine AST substrate (#31417) 2026-06-04 08:27:01 +00:00
containers feat(core): Use n8n default sandbox for Instance AI (no-changelog) (#31335) 2026-06-04 08:31:27 +00:00
janitor refactor: Extract @n8n/test-impact (TIA) package (no-changelog) (#31868) 2026-06-08 10:20:25 +00:00
performance chore: Standardize license metadata across all first-party packages and tighten SBOM pipeline (no-changelog) (#31880) 2026-06-08 10:19:32 +00:00
playwright fix(coverage): Exclude test files, mocks, and .d.ts from coverage collection (no-changelog) 2026-06-08 20:30:39 +01:00
rules-engine refactor: Consolidate janitor + code-health onto a shared rules-engine AST substrate (#31417) 2026-06-04 08:27:01 +00:00
test-impact refactor: Extract @n8n/test-impact (TIA) package (no-changelog) (#31868) 2026-06-08 10:20:25 +00:00
README.md refactor: Extract @n8n/test-impact (TIA) package (no-changelog) (#31868) 2026-06-08 10:20:25 +00:00

packages/testing

n8n's test platform — the infrastructure that decides what to test, runs it, and measures it. Each package is one concern; they compose rather than overlap.

Packages

Package Name Purpose Entry point Consumed by
rules-engine @n8n/rules-engine Generic, typed rules engine for static-analysis tools (register → run → report). The shared substrate. library janitor, code-health
test-impact @n8n/test-impact Test Impact Analysis: build the coverage→impact map, select impacted specs, distribute them across shards. Framework-agnostic. library janitor (CLI), playwright
janitor @n8n/playwright-janitor Static analysis + architecture enforcement for the Playwright suite; also hosts the impact/orchestrate CLI used by CI. janitor CLI playwright, CI
code-health @n8n/code-health Static analysis for monorepo dependency hygiene. code-health CLI CI
containers n8n-containers Composable Docker stack for tests (sqlite / postgres / queue / multi-main / observability / kafka …). stack:* scripts playwright, local dev
playwright n8n-playwright The E2E harness — page objects, composables, fixtures, and the shard distributor that drives CI. test:* scripts CI, local dev
performance @n8n/performance Microbenchmarks for critical code paths (bench, baseline, compare). bench:* scripts CI (nightly), local

How they fit together

graph TD
  rules[rules-engine<br/>rules substrate] --> janitor[janitor<br/>PW lint + impact CLI]
  rules --> codehealth[code-health<br/>monorepo deps]
  impact[test-impact<br/>select + distribute] --> janitor
  impact --> playwright[playwright<br/>E2E harness]
  janitor --> playwright
  containers[containers<br/>Docker stacks] --> playwright
  performance[performance<br/>benchmarks]

  classDef substrate fill:#eef,stroke:#88a;
  class rules,impact substrate;
  • Static-analysis lane: rules-engine (substrate) → janitor (Playwright architecture) and code-health (monorepo deps).
  • Selection lane: test-impact (coverage map → select → distribute) feeds the janitor CLI and the playwright shard distributor.
  • Execution lane: playwright runs E2E against containers-provided stacks.
  • Performance lane: performance is standalone (benchmarks).

Where to look (quick nav for humans + agents)

I want to… Go to
change which E2E specs a PR runs (impact map, selection, sharding) test-impact/ (+ janitor CLI select / distribute)
add/modify an architecture-lint rule for tests janitor/src/rules/ (engine: rules-engine/)
write or fix an E2E test, page object, or fixture playwright/ (see its AGENTS.md)
spin a DB/queue/multi-main stack for a test containers/ (stack:*)
add a microbenchmark performance/
enforce a monorepo dependency rule code-health/

Per-package detail lives in each package's own README.md (and playwright/AGENTS.md).