windsurf-load-scale

# Windsurf Load & Scale ## Overview Strategies for deploying Windsurf AI IDE across large organizations (50-1000+ developers). Covers workspace partitioning for monorepos, configuration distribution, credit budgeting, and performance at scale. ## Prerequisites - Windsurf Teams or Enterprise plan - Admin dashboard access - Understanding of team structure and repository layout - Network/IT involvement for enterprise features ## Instructions ### Step 1: Workspace Strategy for Large Codebases ```yaml # Windsurf performance degrades with workspace size # Cascade context quality inversely correlates with file count workspace_sizing: optimal: "<5,000 files — fast indexing, precise Cascade context" acceptable: "5,000-20,000 files — add .codeiumignore, expect slower indexing" problematic: "20,000+ files — must partition into sub-workspaces" unworkable: "100,000+ files at root — Cascade context diluted, indexing very slow" # Strategy: one Windsurf window per service/package # Each developer opens their assigned service directory ``` ### Step 2: Monorepo Partitioning ``` # Large monorepo (100K+ files) company-monorepo/ ├── .windsurfrules # Brief shared conventions only ├── .codeiumignore # Aggressive: exclude EVERYTHING except src ├── apps/ │ ├── web-app/ # Developer A opens this window │ │ ├── .windsurfrules # Next.js-specific AI context │ │ └── .codeiumignore # Local exclusions │ ├── mobile-app/ ...