miro-performance-tuning

# Miro Performance Tuning ## Overview Optimize Miro REST API v2 throughput and latency. Key levers: minimize API calls with cursor pagination, cache board/item data, batch writes with controlled concurrency, and use connection pooling. ## Latency Benchmarks Typical latencies for `api.miro.com` (US region): | Operation | Endpoint | P50 | P95 | Credits | |-----------|----------|-----|-----|---------| | Get board | `GET /v2/boards/{id}` | 80ms | 200ms | Level 1 | | List items (50) | `GET /v2/boards/{id}/items?limit=50` | 120ms | 350ms | Level 1 | | Create sticky note | `POST /v2/boards/{id}/sticky_notes` | 150ms | 400ms | Level 2 | | Create connector | `POST /v2/boards/{id}/connectors` | 160ms | 420ms | Level 2 | | Update item | `PATCH /v2/boards/{id}/items/{id}` | 130ms | 350ms | Level 2 | | Delete item | `DELETE /v2/boards/{id}/items/{id}` | 100ms | 280ms | Level 2 | ## Cursor Pagination (Eliminate Over-Fetching) Miro v2 uses cursor-based pagination. Fetch only what you need. ```typescript // Efficient paginated iterator async function* paginateItems( boardId: string, options: { type?: string; limit?: number } = {} ): AsyncGenerator<MiroBoardItem> { const limit = options.limit ?? 50; // Max 50 per page let cursor: string | undefined; do { const params = new URLSearchParams({ limit: String(limit) }); if (options.type) params.set('type', options.type); if (cursor) params.set('cursor', cursor); const response = await fetch( `https://api....