Why I Stopped Using Docker for Local Development

Controversial take: Docker is amazing for deployment. But for local development, it's slowing you down. I used Docker Compose for everything — Postgres, Redis, my app, even Node for frontend builds. Then I measured the cost. The Numbers Task Without Docker With Docker Start dev environment 2 seconds 45 seconds Hot reload (Python) Instant 3-8 seconds Run test suite 12 seconds 38 seconds File system operations Native speed 2-10x slower (macOS) RAM usage idle 200MB 2.4GB On macOS, Docker's file system layer (VirtioFS/gRPC FUSE) adds measurable latency to every file read. For a Python project with 500+ files, this means every import, every test discovery, every lint run is slower. What I Use Instead For databases: Native installs or managed services # Postgres — native is fast and simple brew install postgresql@16 brew services start postgresql@16 # Redis brew install redis brew services start redis For team consistency, I use mise (formerly rtx) to pin versions: # .mise.toml [tools] pytho