git.stella-ops.org/tests/load/ttfs-load-test.js

/**
 * TTFS (Time to First Signal) Load Test Suite
 * Reference: SPRINT_0341_0001_0001 Task T13
 * 
 * Tests the /first-signal endpoint under various load scenarios.
 * Requirements from Advisory §12.4:
 * - Cache-hit P95 ≤ 250ms
 * - Cold-path P95 ≤ 500ms
 * - Error rate < 0.1%
 */

import http from 'k6/http';
import { check, sleep } from 'k6';
import { Rate, Trend } from 'k6/metrics';
import { textSummary } from 'https://jslib.k6.io/k6-summary/0.0.3/index.js';

// Custom metrics
const cacheHitLatency = new Trend('ttfs_cache_hit_latency_ms');
const coldPathLatency = new Trend('ttfs_cold_path_latency_ms');
const errorRate = new Rate('ttfs_error_rate');
const signalKindCounter = new Rate('ttfs_signal_kind_distribution');

// Configuration
export const options = {
  scenarios: {
    // Scenario 1: Sustained load - simulates normal operation
    sustained: {
      executor: 'constant-arrival-rate',
      rate: 50,
      timeUnit: '1s',
      duration: '5m',
      preAllocatedVUs: 50,
      maxVUs: 100,
      tags: { scenario: 'sustained' },
    },
    // Scenario 2: Spike test - simulates CI pipeline burst
    spike: {
      executor: 'ramping-arrival-rate',
      startRate: 50,
      timeUnit: '1s',
      stages: [
        { duration: '30s', target: 200 },  // Ramp to 200 RPS
        { duration: '1m', target: 200 },   // Hold
        { duration: '30s', target: 50 },   // Ramp down
      ],
      preAllocatedVUs: 100,
      maxVUs: 300,
      startTime: '5m30s',
      tags: { scenario: 'spike' },
    },
    // Scenario 3: Soak test - long running stability
    soak: {
      executor: 'constant-arrival-rate',
      rate: 25,
      timeUnit: '1s',
      duration: '15m',
      preAllocatedVUs: 30,
      maxVUs: 50,
      startTime: '8m',
      tags: { scenario: 'soak' },
    },
  },
  thresholds: {
    // Advisory requirements: §12.4
    'ttfs_cache_hit_latency_ms{scenario:sustained}': ['p(95)<250'],  // P95 ≤ 250ms
    'ttfs_cache_hit_latency_ms{scenario:spike}': ['p(95)<350'],      // Allow slightly higher during spike
    'ttfs_cold_path_latency_ms{scenario:sustained}': ['p(95)<500'],  // P95 ≤ 500ms
    'ttfs_cold_path_latency_ms{scenario:spike}': ['p(95)<750'],      // Allow slightly higher during spike
    'ttfs_error_rate': ['rate<0.001'],                               // < 0.1% errors
    'http_req_duration{scenario:sustained}': ['p(95)<300'],
    'http_req_duration{scenario:spike}': ['p(95)<500'],
    'http_req_failed': ['rate<0.01'],                                // HTTP failures < 1%
  },
};

// Environment configuration
const BASE_URL = __ENV.BASE_URL || 'http://localhost:5000';
const RUN_IDS = JSON.parse(__ENV.RUN_IDS || '["run-load-1","run-load-2","run-load-3","run-load-4","run-load-5"]');
const TENANT_ID = __ENV.TENANT_ID || 'load-test-tenant';
const AUTH_TOKEN = __ENV.AUTH_TOKEN || '';

/**
 * Main test function - called for each VU iteration
 */
export default function () {
  const runId = RUN_IDS[Math.floor(Math.random() * RUN_IDS.length)];
  const url = `${BASE_URL}/api/v1/orchestrator/runs/${runId}/first-signal`;

  const params = {
    headers: {
      'Accept': 'application/json',
      'X-Tenant-Id': TENANT_ID,
      'X-Correlation-Id': `load-test-${Date.now()}-${Math.random().toString(36).substr(2, 9)}`,
    },
    tags: { endpoint: 'first-signal' },
  };

  // Add auth if provided
  if (AUTH_TOKEN) {
    params.headers['Authorization'] = `Bearer ${AUTH_TOKEN}`;
  }

  const start = Date.now();
  const response = http.get(url, params);
  const duration = Date.now() - start;

  // Track latency by cache status
  const cacheStatus = response.headers['Cache-Status'] || response.headers['X-Cache-Status'];
  if (cacheStatus && cacheStatus.toLowerCase().includes('hit')) {
    cacheHitLatency.add(duration);
  } else {
    coldPathLatency.add(duration);
  }

  // Validate response
  const checks = check(response, {
    'status is 200 or 204 or 304': (r) => [200, 204, 304].includes(r.status),
    'has ETag header': (r) => r.status === 200 ? !!r.headers['ETag'] : true,
    'has Cache-Status header': (r) => !!cacheStatus,
    'response time < 500ms': (r) => r.timings.duration < 500,
    'valid JSON response': (r) => {
      if (r.status !== 200) return true;
      try {
        const body = JSON.parse(r.body);
        return body.runId !== undefined;
      } catch {
        return false;
      }
    },
    'has signal kind': (r) => {
      if (r.status !== 200) return true;
      try {
        const body = JSON.parse(r.body);
        return !body.firstSignal || ['passed', 'failed', 'degraded', 'partial', 'pending'].includes(body.firstSignal.kind);
      } catch {
        return false;
      }
    },
  });

  errorRate.add(!checks);

  // Extract signal kind for distribution analysis
  if (response.status === 200) {
    try {
      const body = JSON.parse(response.body);
      if (body.firstSignal?.kind) {
        signalKindCounter.add(1, { kind: body.firstSignal.kind });
      }
    } catch {
      // Ignore parse errors
    }
  }

  // Minimal sleep to allow for realistic load patterns
  sleep(0.05 + Math.random() * 0.1); // 50-150ms between requests per VU
}

/**
 * Conditional request test - tests ETag/304 behavior
 */
export function conditionalRequest() {
  const runId = RUN_IDS[0];
  const url = `${BASE_URL}/api/v1/orchestrator/runs/${runId}/first-signal`;

  // First request to get ETag
  const firstResponse = http.get(url, {
    headers: { 'Accept': 'application/json', 'X-Tenant-Id': TENANT_ID },
  });

  if (firstResponse.status !== 200) return;

  const etag = firstResponse.headers['ETag'];
  if (!etag) return;

  // Conditional request
  const conditionalResponse = http.get(url, {
    headers: {
      'Accept': 'application/json',
      'X-Tenant-Id': TENANT_ID,
      'If-None-Match': etag,
    },
    tags: { request_type: 'conditional' },
  });

  check(conditionalResponse, {
    'conditional request returns 304': (r) => r.status === 304,
  });
}

/**
 * Setup function - runs once before the test
 */
export function setup() {
  console.log(`Starting TTFS load test against ${BASE_URL}`);
  console.log(`Testing with ${RUN_IDS.length} run IDs`);
  
  // Verify endpoint is accessible
  const healthCheck = http.get(`${BASE_URL}/health`, { timeout: '5s' });
  if (healthCheck.status !== 200) {
    console.warn(`Health check returned ${healthCheck.status} - proceeding anyway`);
  }

  return { startTime: Date.now() };
}

/**
 * Teardown function - runs once after the test
 */
export function teardown(data) {
  const duration = (Date.now() - data.startTime) / 1000;
  console.log(`TTFS load test completed in ${duration.toFixed(1)}s`);
}

/**
 * Generate test summary
 */
export function handleSummary(data) {
  const timestamp = new Date().toISOString().replace(/[:.]/g, '-');
  
  return {
    'stdout': textSummary(data, { indent: ' ', enableColors: true }),
    [`results/ttfs-load-test-${timestamp}.json`]: JSON.stringify(data, null, 2),
    'results/ttfs-load-test-latest.json': JSON.stringify(data, null, 2),
  };
}