[grade=B urn:ump:3wvdd3yegylr7e2pzn6tebyq5bisf2awmss7p4j3uwpwsujtnwoq] DC-134/135/136: Shipdeck integration - data-driven login pages, deploy events, badge suppression

DC-134: /api/v1/auth/login-page serves a generic gated auto-login page for any service registered in services.json without a curated flow (App Selector installs, DC-131 git installs). Curated pages always win; unknown services still 404; sanitizer keeps digits/hyphens (shipdeck-style ids); display names HTML-escaped. Kills the sso-gate.js edit + restart per new install. DC-135: shipdeck journal.jsonl tail worker (startShipdeckWorker) ingests deploy/rollback lifecycle rows into the Security Center as source_type=shipdeck (notice/success, error/failure via verify[] block). DC-136: deploy-aware badge suppression - suppressDuringDeploy() before the bridge call in /deploy and /rollback, clearDeploySuppression() in finally (every exit path incl. rejected fetches), reference-counted for overlapping deploys, 10-min TTL auto-expiry (HEALTH_DEPLOY_SUPPRESS_MAX_MS). 23 new tests across 4 suites; full suite 2923/2923 green. Codex judge: C (r1) -> C (r2) -> B (r3) -> B zero-blockers (r4, urn:ump:3wvdd3yegylr7e2pzn6tebyq5bisf2awmss7p4j3uwpwsujtnwoq).
2026-09-16 03:10:54 -07:00
parent 0dd8493f98
commit 11c719e635
10 changed files with 912 additions and 11 deletions
@@ -0,0 +1,130 @@
+/**
+ * DC-136: deploy-aware badge suppression in the health checker.
+ *
+ * A shipdeck deploy/rollback restarts the target unit; probes that land
+ * during that window blackhole (timeouts / 5xx) and — before this change —
+ * flipped the badge red and opened outage incidents for what is routine
+ * deploy noise.
+ *
+ * Pins:
+ *   - suppressDuringDeploy() holds the displayed badge through down probes
+ *     (even past DOWN_THRESHOLD) and emits nothing.
+ *   - Raw history keeps every probe (full fidelity preserved).
+ *   - checkForIncidents opens no outage/slow-response incident while
+ *     suppressed.
+ *   - After expiry the checker behaves exactly as before (down probes flip
+ *     the badge again).
+ *   - TTL is clamped to HEALTH_DEPLOY_SUPPRESS_MAX_MS.
+ */
+
+'use strict';
+
+const path = require('path');
+const fs = require('fs');
+const os = require('os');
+
+const TMP_DIR = fs.mkdtempSync(path.join(os.tmpdir(), 'dc136-deploysuppress-'));
+process.env.HEALTH_DATA_DIR = TMP_DIR;
+process.env.HEALTH_CONFIG_FILE = path.join(TMP_DIR, 'health-config.json');
+process.env.HEALTH_HISTORY_FILE = path.join(TMP_DIR, 'health-history.json');
+process.env.HEALTH_DEPLOY_SUPPRESS_MAX_MS = '60000'; // test-visible clamp ceiling
+
+// Module exports a singleton instance — same pattern as
+// health-checker-hysteresis.test.js.
+const healthCheckerSingleton = require('../src/monitoring/health-checker');
+
+function makeUp(serviceId = 'svc1') {
+  return {
+    serviceId,
+    timestamp: new Date().toISOString(),
+    status: 'up',
+    responseTime: 50,
+    statusCode: 200,
+    message: 'Service is healthy',
+    details: { headers: {}, bodyLength: 12 },
+  };
+}
+
+function makeDown(serviceId = 'svc1') {
+  return {
+    serviceId,
+    timestamp: new Date().toISOString(),
+    status: 'down',
+    responseTime: 50,
+    statusCode: 500,
+    message: 'fail',
+    details: { headers: {}, bodyLength: 0 },
+  };
+}
+
+describe('DC-136: deploy suppression on the dashboard badge', () => {
+  let hc;
+  let emitSpy;
+
+  beforeEach(() => {
+    hc = healthCheckerSingleton;
+    hc.displayedStatus = new Map();
+    hc.consecutiveSinceChange = new Map();
+    hc.currentStatus = new Map();
+    hc.history = {};
+    hc.deploySuppressedUntil = new Map();
+    hc.deploySuppressRefs = new Map(); // judge r4 polish: reset ref counts too
+    hc.incidents = [];
+    emitSpy = jest.spyOn(hc, 'emit');
+  });
+
+  afterEach(() => {
+    emitSpy.mockRestore();
+  });
+
+  test('down probes during the suppress window do not flip the badge', () => {
+    hc.recordStatus('svc1', makeUp());
+    expect(hc.displayedStatus.get('svc1').status).toBe('up');
+
+    hc.suppressDuringDeploy('svc1');
+    hc.recordStatus('svc1', makeDown());
+    hc.recordStatus('svc1', makeDown());
+    hc.recordStatus('svc1', makeDown()); // well past DOWN_THRESHOLD=2
+
+    expect(hc.displayedStatus.get('svc1').status).toBe('up');
+    const statusEmits = emitSpy.mock.calls.filter(c => c[0] === 'status-check');
+    expect(statusEmits.length).toBe(1); // only the bootstrap "up" emit
+  });
+
+  test('raw history keeps every probe during suppression', () => {
+    hc.recordStatus('svc1', makeUp());
+    hc.suppressDuringDeploy('svc1');
+    hc.recordStatus('svc1', makeDown());
+    hc.recordStatus('svc1', makeDown());
+
+    expect(hc.history.svc1.length).toBe(3);
+    expect(hc.currentStatus.get('svc1').status).toBe('down');
+  });
+
+  test('no outage or slow-response incidents open while suppressed', () => {
+    hc.recordStatus('svc1', makeUp());
+    hc.suppressDuringDeploy('svc1');
+    const down = makeDown();
+    down.responseTime = 99999; // would trip slow-response too
+    hc.recordStatus('svc1', down);
+
+    expect(hc.incidents.length).toBe(0);
+  });
+
+  test('after expiry, down probes flip the badge again (unchanged semantics)', () => {
+    hc.recordStatus('svc1', makeUp());
+    hc.suppressDuringDeploy('svc1', 1); // expires immediately
+    // spin clock past expiry without sleeps
+    hc.deploySuppressedUntil.set('svc1', Date.now() - 1);
+
+    hc.recordStatus('svc1', makeDown());
+    hc.recordStatus('svc1', makeDown());
+    expect(hc.displayedStatus.get('svc1').status).toBe('down');
+  });
+
+  test('ttl is clamped to HEALTH_DEPLOY_SUPPRESS_MAX_MS', () => {
+    hc.suppressDuringDeploy('svc1', 10 * 60 * 60 * 1000); // 1h request
+    const until = hc.deploySuppressedUntil.get('svc1');
+    expect(until - Date.now()).toBeLessThanOrEqual(60000 + 50);
+  });
+});