Scan a Monorepo

Strategies for scanning large codebases efficiently.

The Challenge

Monorepos can have thousands of files. A full scan may take minutes and produce a huge graph.

Strategy 1: Scope to Directories

Scan only relevant directories:

jnkn scan --dir src/services --dir terraform/

Or use .jnknignore:

# .jnknignore
node_modules/
vendor/
*.test.py
*_test.go
__pycache__/
.git/
docs/
scripts/

Strategy 2: Incremental Scanning

Only re-scan changed files:

# First scan (full)
jnkn scan

# Subsequent scans (incremental)
jnkn scan  # Automatically detects unchanged files via hash

Force full rescan if needed:

jnkn scan --full

Strategy 3: Parallel Scanning

Use multiple cores:

jnkn scan --jobs 8

Or auto-detect:

jnkn scan --jobs auto  # Uses CPU count

Strategy 4: Split by Domain

For very large monorepos, maintain separate databases:

# Backend team
jnkn scan --dir backend/ --db .jnkn/backend.db

# Infrastructure team  
jnkn scan --dir terraform/ --db .jnkn/infra.db

# Query across both
jnkn blast env:DATABASE_URL \
  --db .jnkn/backend.db \
  --db .jnkn/infra.db

Strategy 5: CI Caching

Cache the database between CI runs:

- uses: actions/cache@v4
  with:
    path: .jnkn/
    key: jnkn-${{ hashFiles('**/*.py', '**/*.tf') }}
    restore-keys: jnkn-

Performance Tips

Files	Expected Scan Time	Recommendation
< 100	< 5 seconds	Default settings
100-1000	5-30 seconds	Enable caching
1000-10000	30s - 3 min	Parallel + incremental
> 10000	3+ min	Split by domain

Monitoring Scan Performance

jnkn scan --verbose

Output includes timing:

Parsing: 12.3s (847 files)
Stitching: 3.2s (156 rules evaluated)
Storage: 0.8s (2341 nodes written)
Total: 16.3s