exa-load-scale

# Exa Load & Scale ## Overview Load testing and capacity planning for Exa integrations. Key constraint: Exa's default rate limit is 10 QPS. Scaling strategies focus on caching, request queuing, parallel processing within rate limits, and search type selection for latency budgets. ## Prerequisites - k6 load testing tool installed - Test environment Exa API key (separate from production) - Redis for result caching ## Capacity Reference | Search Type | Typical Latency | Max Throughput (10 QPS) | |-------------|----------------|-------------------------| | `instant` | < 150ms | 10 req/s (600/min) | | `fast` | < 425ms | 10 req/s (600/min) | | `auto` | 300-1500ms | 10 req/s (600/min) | | `neural` | 500-2000ms | 10 req/s (600/min) | | `deep` | 2-5s | 10 req/s (600/min) | **With caching (50% hit rate):** Effective throughput doubles to 20 req/s equivalent. ## Instructions ### Step 1: k6 Load Test Against Your Wrapper ```javascript // exa-load-test.js import http from "k6/http"; import { check, sleep } from "k6"; export const options = { stages: [ { duration: "1m", target: 5 }, // Ramp up to 5 VUs { duration: "3m", target: 5 }, // Steady state { duration: "1m", target: 10 }, // Push toward rate limit { duration: "2m", target: 10 }, // Stress test { duration: "1m", target: 0 }, // Ramp down ], thresholds: { http_req_duration: ["p(95)<3000"], // 3s P95 for neural search http_req_failed: ["rate<0.05"], // < 5% error rate }, };...