[glm-grade=A] fix(monitoring): DC-090 outage incidents follow displayed hysteresis status

checkForIncidents compared raw probe transitions while the dashboard badge
(DC-086) follows post-hysteresis displayed status. A single raw down blip
between two ups opened AND resolved a critical outage incident; a suppressed
up blip during a real outage resolved it early. Incidents now open/resolve
on displayed-vs-displayed transitions; previousDisplayed=null keeps legacy
raw semantics for direct callers. 6 new parity tests + legacy checkService
test moved to a 4-probe chain. Suite 2616/2616 (110).

Verdict: urn:ump:yc5rdlnmnmhch5audc5fifgbt6d7moi2stqfs5vidsbh6x6zkvgq

BACKLOG.md

@@ -401,6 +401,22 @@ Tickets DC-046 through DC-049 implement pluggable auth + email magic link. Sami
 - **result:** Shipped codex-graded A. All 49 `console.*` sites in `src/managers/update-manager.js` now route through `log.info/log.warn/log.error` from `src/utils/logging` (tag = `'update'`). Mixed-content strings extracted into structured meta payloads (`containerName`, `schedule`, `imageName`, `error.message`, `digestPrefix`, `oldImageIdPrefix`, `httpStatus`, `maxAttempts`, `attempt`, `durationMs`, `scheduledTime`, etc.) so fields are queryable instead of inlined into the message. Errors now go through `log.error(ctx, errObj)` so they land in error.log with full stack trace + context, not just stderr. 1539/1539 Jest tests pass (78/78 update-manager tests still pass). ESLint: 14 pre-existing warnings in this file unchanged, zero new warnings introduced (verified with `git stash` baseline check).
 
 
+### DC-086: Service-status flicker fix — asymmetric hysteresis on the badge
+- **status:** in-progress
+- **owner:** hermes
+- **details:** Dashboard service badges perpetually flip between green and red for "a few seconds at a time, never stable" (Sami's report, 2026-08-20). Root cause: `src/monitoring/health-checker.js` `recordStatus()` emits `'status-check'` on EVERY probe (every 30s), and `src/websocket/dashboard-ws.js` forwards every probe as `'status-change'` to the browser with no diff. The frontend `live-events.js` then unconditionally calls `setBadge()` — which resets the icon + pill text on every event. A single transient 5xx (Caddy reload, container CPU steal, mid-flight TLS handshake, container restart during probe) flips the badge red and the next green probe flips it back. Fix: add asymmetric hysteresis in `_computeDisplayedStatus(serviceId, rawStatus)` — going DOWN requires 2 consecutive "down" probes (default `HEALTH_DOWN_THRESHOLD=2`), going UP requires only 1 (default `HEALTH_UP_THRESHOLD=1`). History + `consecutiveFailures` still record raw probe results (operators want full fidelity for postmortems); only the dashboard broadcast is filtered. `getCurrentStatus()` now returns the displayed status so a page reload shows the same badge as the live SSE stream. Both thresholds are env-var configurable so operators can tune. New tests in `__tests__/health-checker-hysteresis.test.js` cover: first probe emits; second probe same-status does NOT re-emit; one-down-then-up keeps green; two-down flips to red; one-up after down flips back to green; `getCurrentStatus` returns displayed not raw. Effort: ~30 min. Risk: low — pure behavior filter, no schema breaks, all 63 existing health-checker tests must stay green.
+- **impact:** Operators stop seeing perpetual red/green flicker on healthy services. Real outages still get flagged (2 consecutive 30s probes = ~60s before badge flips red, which is still faster than a human notices). Background probe history is unchanged so postmortem analysis still works.
+- **prerequisite:** None.
+- **result:** _pending — ship + codex round_
+
+### DC-085: Link-first invite — Discord-style "share it however you want"
+- **status:** in-progress
+- **owner:** hermes
+- **details:** Today `POST /api/v1/auth/admin/invites` defaults to sending the invite link via SMTP; if SMTP is not configured it spams the server console with `[DC-048-DEV-INVITE-LINK]` log lines. Sami wants Discord-style: the link is always returned in the response, and email is an opt-in checkbox. Operators should be free to copy the link and share it via iMessage / SMS / WhatsApp / Telegram / Signal / Discord / paste-in-email — whatever fits. (1) Flip default `sendEmail !== false` to `sendEmail === true` in `routes/auth/admin.js` so omitting the field means "no email, just hand me the link." (2) Stop logging the raw invite URL to error.log when SMTP is unconfigured — that path was only useful when there was no UI way to grab the link; now there is. (3) Add a `shareText` field to the response: `"Join my DashCaddy as <role> — <acceptUrl> — expires in Nh."` for one-tap paste into any messenger. (4) Frontend: `status/js/admin.js` `_renderInviteForm` flips the "Send email" checkbox default to **unchecked**, updates `_renderIssuedInviteBanner` to show both the raw link AND the shareText (with its own copy button + `navigator.share()` native share-sheet button where available). (5) New tests in `__tests__/admin-invites.test.js` covering: default sendEmail=false (no SMTP send attempted, no console log); `sendEmail: true` triggers SMTP send; `shareText` is present and well-formed; `acceptUrl` is always returned; expired sendEmail path doesn't leak token to logs. Effort: ~1 hr. Risk: low — pure behavior flip + UI additive change.
+- **impact:** Closes the friction between "host wants to add a friend" and "host has to configure SMTP first." Mirrors Discord/Slack/Linear invite flows where the link IS the deliverable. No new tier changes, no schema breaks.
+- **prerequisite:** DC-048 (invite store + admin route), DC-052 (Pro gate stays).
+- **result:** _pending — ship + codex round_
+
 ### DC-084: Remove redundant active Caddy health check from `arch.sami` site — eliminate 6 syslog spam lines/min
 - **status:** done
 - **owner:** hermes

dashcaddy-api/__tests__/admin-invites.test.js

@@ -0,0 +1,285 @@
+/**
+ * Tests for DC-085: link-first invite (Discord-style "share it however you want").
+ *
+ *   - default sendEmail omission = no email sent, link returned, no token in logs
+ *   - sendEmail:true triggers SMTP send when configured
+ *   - sendEmail:true + SMTP unconfigured = deliveredVia:'failed', no token leaked
+ *   - shareText field present and well-formed in every response
+ *   - acceptUrl always present (regardless of sendEmail)
+ *   - role + ttl validation unchanged from DC-048
+ *
+ * Strategy: drive the route handler directly with mock req/res, mount the admin
+ * router against an isolated userStore + inviteStore + email-sender stub.
+ */
+
+'use strict';
+
+const fs = require('fs');
+const path = require('path');
+const os = require('os');
+const express = require('express');
+
+function _tmpDir() {
+  return fs.mkdtempSync(path.join(os.tmpdir(), 'dashcaddy-admin-invites-test-'));
+}
+function _cleanup(dir) {
+  try { fs.rmSync(dir, { recursive: true, force: true }); } catch {}
+}
+
+// Stub email-sender so we can assert "was it called?" without an SMTP server.
+// NOTE: the variable name MUST start with `mock` so Jest's hoisted `jest.mock()`
+// call is allowed to reference it (Babel guard against out-of-scope access).
+const mockEmailSender = {
+  isConfigured: jest.fn(() => false),
+  sendEmail: jest.fn(async () => undefined),
+};
+jest.mock('../src/auth/providers/email-sender', () => mockEmailSender);
+
+describe('DC-085: link-first admin invites', () => {
+  let dir, app, request;
+  let logCalls; // captured { level, msg, meta } from our fake log
+
+  beforeEach(async () => {
+    jest.clearAllMocks();
+    dir = _tmpDir();
+    logCalls = [];
+
+    // Set up email auth enable flag so userStore mounts.
+    process.env.NODE_ENV = 'test';
+
+    const { createUserStore } = require('../src/security/user-store');
+    const userStore = createUserStore({ dataDir: dir });
+
+    // Bootstrap the admin so we have a session-attributable user.
+    await userStore.login({ email: 'admin@sami-host.me' });
+
+    // Build a tiny Express app with the admin router mounted, but skip the
+    // global auth gate (we inject req.user directly).
+    const adminRouter = require('../routes/auth/admin')({
+      asyncHandler: (fn) => (req, res, next) => Promise.resolve(fn(req, res, next)).catch(next),
+      errorResponse: (_res, code, msg) => ({ status: code, msg }),
+      log: {
+        info: (topic, msg, meta) => logCalls.push({ level: 'info', topic, msg, meta }),
+        warn: (topic, msg, meta) => logCalls.push({ level: 'warn', topic, msg, meta }),
+        error: (topic, msg, meta) => logCalls.push({ level: 'error', topic, msg, meta }),
+      },
+      session: { isSessionValid: () => true, create: () => {}, setCookie: () => {} },
+      dataDir: dir,
+    });
+
+    app = express();
+    app.use(express.json());
+    // Inject req.user = admin so /admin/* passes the role gate.
+    app.use((req, _res, next) => {
+      req.user = { id: 'admin-id', email: 'admin@sami-host.me', role: 'admin' };
+      req.app.locals = req.app.locals || {};
+      req.app.locals.siteConfig = {}; // no publicBaseUrl — route uses req.headers
+      req.app.locals.emailConfig = null; // SMTP not configured by default
+      next();
+    });
+    app.use('/api/v1/auth', adminRouter);
+    // Error handler — last in chain.
+    app.use((err, _req, res, _next) => {
+      const code = (err && err.statusCode) || 500;
+      res.status(code).json({
+        success: false,
+        error: err && err.message,
+        code: err && err.code,
+      });
+    });
+
+    request = require('supertest');
+  });
+
+  afterEach(() => _cleanup(dir));
+
+  test('default sendEmail (omitted) returns link and does NOT send email', async () => {
+    const res = await request(app)
+      .post('/api/v1/auth/admin/invites')
+      .send({ email: 'friend@example.com', role: 'operator' });
+
+    expect(res.status).toBe(200);
+    expect(res.body.success).toBe(true);
+    expect(mockEmailSender.sendEmail).not.toHaveBeenCalled();
+    expect(res.body.acceptUrl).toMatch(/\/api\/v1\/auth\/invites\/[^/]+\/accept$/);
+    expect(res.body.deliveredVia).toBe('manual');
+  });
+
+  test('default sendEmail does NOT log raw token to server log', async () => {
+    const res = await request(app)
+      .post('/api/v1/auth/admin/invites')
+      .send({ email: 'friend@example.com', role: 'operator' });
+
+    const acceptUrl = res.body.acceptUrl;
+    // Extract the token from the URL and verify it does NOT appear in any log call.
+    const token = acceptUrl.match(/invites\/([^/]+)\/accept/)[1];
+    const tokenLeaked = logCalls.some(c =>
+      typeof c.msg === 'string' && c.msg.includes(token)
+    );
+    expect(tokenLeaked).toBe(false);
+
+    // Also assert no log entry mentions the URL verbatim (the old
+    // `[DC-048-DEV-INVITE-LINK] url=...` spam).
+    const oldSpam = logCalls.find(c =>
+      typeof c.msg === 'string' && c.msg.includes('[DC-048-DEV-INVITE-LINK]')
+    );
+    expect(oldSpam).toBeUndefined();
+  });
+
+  test('shareText is present and well-formed in every response', async () => {
+    const res = await request(app)
+      .post('/api/v1/auth/admin/invites')
+      .send({ email: 'friend@example.com', role: 'operator', ttlHours: 24 });
+
+    expect(res.body.shareText).toBeDefined();
+    expect(res.body.shareText).toContain('Join my DashCaddy');
+    expect(res.body.shareText).toContain('operator');
+    expect(res.body.shareText).toContain(res.body.acceptUrl);
+    expect(res.body.shareText).toContain('expires in 24h');
+  });
+
+  test('acceptUrl is always returned regardless of sendEmail', async () => {
+    const r1 = await request(app)
+      .post('/api/v1/auth/admin/invites')
+      .send({ email: 'a@x.com', sendEmail: false });
+    const r2 = await request(app)
+      .post('/api/v1/auth/admin/invites')
+      .send({ email: 'b@x.com' });
+    expect(r1.body.acceptUrl).toBeTruthy();
+    expect(r2.body.acceptUrl).toBeTruthy();
+  });
+
+  test('sendEmail: true triggers SMTP send when configured', async () => {
+    // Build a SECOND app instance where emailConfig is a real-looking object,
+    // so isConfigured() returns true. The first app uses emailConfig=null.
+    mockEmailSender.isConfigured.mockReturnValueOnce(true);
+    mockEmailSender.sendEmail.mockResolvedValueOnce(undefined);
+    const app2 = express();
+    app2.use(express.json());
+    app2.use((req, _res, next) => {
+      req.user = { id: 'admin-id', email: 'admin@sami-host.me', role: 'admin' };
+      req.app.locals = req.app.locals || {};
+      req.app.locals.siteConfig = {};
+      req.app.locals.emailConfig = { host: 'smtp.test', from: 'noreply@test' };
+      next();
+    });
+    const { createUserStore } = require('../src/security/user-store');
+    const userStore2 = createUserStore({ dataDir: dir });
+    await userStore2.login({ email: 'admin@sami-host.me' });
+    const router2 = require('../routes/auth/admin')({
+      asyncHandler: (fn) => (req, res, next) => Promise.resolve(fn(req, res, next)).catch(next),
+      errorResponse: (_res, code, msg) => ({ status: code, msg }),
+      log: { info() {}, warn: (t, m, meta) => logCalls.push({ level: 'warn', topic: t, msg: m, meta }), error() {} },
+      session: { isSessionValid: () => true, create: () => {}, setCookie: () => {} },
+      dataDir: dir,
+    });
+    app2.use('/api/v1/auth', router2);
+
+    const res = await request(app2)
+      .post('/api/v1/auth/admin/invites')
+      .send({ email: 'friend@example.com', role: 'viewer', sendEmail: true });
+
+    expect(res.status).toBe(200);
+    expect(mockEmailSender.sendEmail).toHaveBeenCalledTimes(1);
+    const [_cfg, to, subject, text, html] = mockEmailSender.sendEmail.mock.calls[0];
+    expect(to).toBe('friend@example.com');
+    expect(subject).toMatch(/invited/i);
+    expect(text).toContain(res.body.acceptUrl);
+    expect(html).toContain(res.body.acceptUrl);
+    expect(res.body.deliveredVia).toBe('email');
+  });
+
+  test('sendEmail: true + SMTP unconfigured returns deliveredVia:failed and does NOT leak token', async () => {
+    mockEmailSender.isConfigured.mockReturnValueOnce(false);
+
+    const res = await request(app)
+      .post('/api/v1/auth/admin/invites')
+      .send({ email: 'friend@example.com', role: 'operator', sendEmail: true });
+
+    expect(res.status).toBe(200);
+    expect(mockEmailSender.sendEmail).not.toHaveBeenCalled();
+    expect(res.body.deliveredVia).toBe('failed');
+    // acceptUrl + shareText still present so the operator can share manually.
+    expect(res.body.acceptUrl).toBeTruthy();
+    expect(res.body.shareText).toBeTruthy();
+    // Token does NOT appear in any log call.
+    const token = res.body.acceptUrl.match(/invites\/([^/]+)\/accept/)[1];
+    const tokenLeaked = logCalls.some(c =>
+      typeof c.msg === 'string' && c.msg.includes(token)
+    );
+    expect(tokenLeaked).toBe(false);
+  });
+
+  test('invalid role silently defaults to operator (DC-048 behavior preserved)', async () => {
+    // DC-048: the route's `(role && VALID_ROLES.has(role)) ? role : 'operator'`
+    // silently substitutes default rather than throwing. This test pins that
+    // behavior so a future "strict role validation" change is a deliberate
+    // decision, not a silent regression.
+    const res = await request(app)
+      .post('/api/v1/auth/admin/invites')
+      .send({ email: 'a@x.com', role: 'superuser' });
+    expect(res.status).toBe(200);
+    expect(res.body.role).toBe('operator');
+    expect(mockEmailSender.sendEmail).not.toHaveBeenCalled();
+  });
+
+  test('email validation: missing email still rejected', async () => {
+    const res = await request(app)
+      .post('/api/v1/auth/admin/invites')
+      .send({ role: 'operator' });
+    expect(res.status).toBe(400);
+    expect(mockEmailSender.sendEmail).not.toHaveBeenCalled();
+  });
+
+  test('ttlHours: 1 still produces shareText with correct expiry wording', async () => {
+    const res = await request(app)
+      .post('/api/v1/auth/admin/invites')
+      .send({ email: 'a@x.com', ttlHours: 1 });
+    expect(res.body.shareText).toContain('expires in 1h');
+  });
+
+  test('DC-089: SMTP-unconfigured warn log masks the invite email (no raw PII)', async () => {
+    mockEmailSender.isConfigured.mockReturnValueOnce(false);
+
+    const res = await request(app)
+      .post('/api/v1/auth/admin/invites')
+      .send({ email: 'friend@example.com', role: 'operator', sendEmail: true });
+
+    expect(res.status).toBe(200);
+    expect(res.body.deliveredVia).toBe('failed');
+    const warn = logCalls.find(c =>
+      c.level === 'warn' && c.topic === 'auth-invite-send'
+    );
+    expect(warn).toBeDefined();
+    // The raw address must not appear; the masked form must.
+    expect(JSON.stringify(warn.meta)).not.toContain('friend@example.com');
+    expect(warn.meta.email).toBe('fr****@example.com');
+  });
+
+  test('DC-089: invite-accepted info log masks the created user email (no raw PII)', async () => {
+    // Pre-authorize the email (POST /admin/users) so userStore.login doesn't
+    // reject with not_authorized — bootstrap already happened in beforeEach.
+    const preauth = await request(app)
+      .post('/api/v1/auth/admin/users')
+      .send({ email: 'newfriend@example.com' });
+    expect(preauth.status).toBe(200);
+
+    const issue = await request(app)
+      .post('/api/v1/auth/admin/invites')
+      .send({ email: 'newfriend@example.com', role: 'viewer' });
+    expect(issue.status).toBe(200);
+    const token = issue.body.acceptUrl.match(/invites\/([^/]+)\/accept/)[1];
+
+    const res = await request(app)
+      .post(`/api/v1/auth/invites/${token}/accept`)
+      .send({});
+
+    expect(res.status).toBe(200);
+    const info = logCalls.find(c =>
+      c.level === 'info' && c.msg === 'invite accepted, user created'
+    );
+    expect(info).toBeDefined();
+    expect(JSON.stringify(info.meta)).not.toContain('newfriend@example.com');
+    expect(info.meta.email).toBe('ne****@example.com');
+  });
+});

dashcaddy-api/__tests__/health-checker-hysteresis.test.js

@@ -0,0 +1,297 @@
+/**
+ * Tests for DC-086: asymmetric hysteresis on the dashboard service badge.
+ *
+ *   - First probe always emits (no prior state).
+ *   - Same-status probe does NOT re-emit (dedup against repeated green).
+ *   - One "down" then back to "up" keeps the badge green (no flicker).
+ *   - Two consecutive "down" probes flip the badge to red.
+ *   - One "up" after a down streak flips back to green (fast recovery).
+ *   - History retains every raw probe even when no emit happens.
+ *   - getCurrentStatus returns displayed status, not raw.
+ */
+
+'use strict';
+
+const path = require('path');
+const fs = require('fs');
+const os = require('os');
+
+// Use an isolated data dir so test history doesn't pollute the real one.
+const tmpDir = fs.mkdtempSync(path.join(os.tmpdir(), 'dashcaddy-hyst-'));
+process.env.HEALTH_DATA_DIR = tmpDir;
+process.env.HEALTH_CONFIG_FILE = path.join(tmpDir, 'health-config.json');
+process.env.HEALTH_HISTORY_FILE = path.join(tmpDir, 'health-history.json');
+
+// Module exports a singleton instance, not a class — see module.exports in
+// src/monitoring/health-checker.js. The test creates fresh state by replacing
+// the relevant maps on the singleton in beforeEach.
+const healthCheckerSingleton = require('../src/monitoring/health-checker');
+const originalDownThreshold = process.env.HEALTH_DOWN_THRESHOLD;
+const originalUpThreshold = process.env.HEALTH_UP_THRESHOLD;
+
+function restoreEnv(name, value) {
+  if (value === undefined) delete process.env[name];
+  else process.env[name] = value;
+}
+
+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-086: hysteresis on the dashboard badge', () => {
+  let hc;
+  let emitSpy;
+
+  beforeEach(() => {
+    // Reset the singleton's per-test state so each case starts clean.
+    healthCheckerSingleton.displayedStatus = new Map();
+    healthCheckerSingleton.consecutiveSinceChange = new Map();
+    healthCheckerSingleton.currentStatus = new Map();
+    healthCheckerSingleton.history = {};
+    healthCheckerSingleton.removeAllListeners('status-check');
+    emitSpy = jest.fn();
+    healthCheckerSingleton.on('status-check', emitSpy);
+    hc = healthCheckerSingleton;
+  });
+
+  afterEach(() => {
+    restoreEnv('HEALTH_DOWN_THRESHOLD', originalDownThreshold);
+    restoreEnv('HEALTH_UP_THRESHOLD', originalUpThreshold);
+  });
+
+  afterAll(() => {
+    fs.rmSync(tmpDir, { recursive: true, force: true });
+  });
+
+  test('first probe (no prior state) emits', () => {
+    hc.recordStatus('svc1', makeUp());
+    expect(emitSpy).toHaveBeenCalledTimes(1);
+    expect(emitSpy.mock.calls[0][0].status).toBe('up');
+  });
+
+  test('second probe with same status does NOT re-emit', () => {
+    hc.recordStatus('svc1', makeUp());
+    hc.recordStatus('svc1', makeUp());
+    expect(emitSpy).toHaveBeenCalledTimes(1);
+  });
+
+  test('one "down" then "up" keeps the badge green (the flicker bug)', () => {
+    hc.recordStatus('svc1', makeUp());        // baseline: green, emit 1
+    hc.recordStatus('svc1', makeDown());      // one blip — keep green, no emit
+    hc.recordStatus('svc1', makeUp());        // recovered — still green, no emit
+    expect(emitSpy).toHaveBeenCalledTimes(1);
+    expect(hc.displayedStatus.get('svc1').status).toBe('up');
+  });
+
+  test('up, down, up, down, down resets the first streak before flipping', () => {
+    hc.recordStatus('svc1', makeUp());
+    hc.recordStatus('svc1', makeDown());
+    hc.recordStatus('svc1', makeUp());
+    hc.recordStatus('svc1', makeDown());
+    expect(hc.displayedStatus.get('svc1').status).toBe('up');
+    expect(emitSpy).toHaveBeenCalledTimes(1);
+
+    hc.recordStatus('svc1', makeDown());
+    expect(hc.displayedStatus.get('svc1').status).toBe('down');
+    expect(emitSpy).toHaveBeenCalledTimes(2);
+  });
+
+  test('two consecutive "down" probes flip the badge to red', () => {
+    hc.recordStatus('svc1', makeUp());        // baseline: green
+    hc.recordStatus('svc1', makeDown());      // blip #1 — keep green (counter=1)
+    hc.recordStatus('svc1', makeDown());      // blip #2 — flip red (counter=2 >= DOWN_THRESHOLD)
+    expect(emitSpy).toHaveBeenCalledTimes(2);
+    expect(emitSpy.mock.calls[1][0].status).toBe('down');
+    expect(hc.displayedStatus.get('svc1').status).toBe('down');
+  });
+
+  test('one "up" after a down streak flips back to green (fast recovery)', () => {
+    hc.recordStatus('svc1', makeUp());
+    hc.recordStatus('svc1', makeDown());
+    hc.recordStatus('svc1', makeDown());      // now red
+    expect(hc.displayedStatus.get('svc1').status).toBe('down');
+
+    hc.recordStatus('svc1', makeUp());        // first green — flip back
+    expect(emitSpy).toHaveBeenCalledTimes(3);
+    expect(emitSpy.mock.calls[2][0].status).toBe('up');
+    expect(hc.displayedStatus.get('svc1').status).toBe('up');
+  });
+
+  test('history retains every raw probe even when no emit happens', () => {
+    hc.recordStatus('svc1', makeUp());
+    hc.recordStatus('svc1', makeDown());      // blip, no emit
+    hc.recordStatus('svc1', makeUp());        // recovery, no emit
+    expect(hc.history['svc1'].length).toBe(3);
+    expect(hc.history['svc1'][0].status).toBe('up');
+    expect(hc.history['svc1'][1].status).toBe('down');
+    expect(hc.history['svc1'][2].status).toBe('up');
+  });
+
+  test('getCurrentStatus returns the displayed status, not the raw probe', () => {
+    const displayedUp = makeUp();
+    displayedUp.timestamp = '2026-08-22T09:59:00.000Z';
+    displayedUp.statusCode = 200;
+    displayedUp.message = 'healthy';
+    displayedUp.details = { source: 'accepted-up' };
+    hc.recordStatus('svc1', displayedUp);
+    const latestRaw = makeDown();
+    latestRaw.timestamp = '2026-08-22T10:00:00.000Z';
+    latestRaw.responseTime = 987;
+    latestRaw.statusCode = 500;
+    latestRaw.message = 'failed probe';
+    latestRaw.error = 'upstream failure';
+    latestRaw.details = { source: 'suppressed-down' };
+    hc.recordStatus('svc1', latestRaw);       // raw=down, displayed=up
+    const out = hc.getCurrentStatus();
+    expect(out['svc1'].status).toBe('up');    // shown to API consumers
+    expect(out['svc1'].timestamp).toBe(displayedUp.timestamp);
+    expect(out['svc1'].statusCode).toBe(200);
+    expect(out['svc1'].message).toBe('healthy');
+    expect(out['svc1'].error).toBeUndefined();
+    expect(out['svc1'].details).toEqual({ source: 'accepted-up' });
+    expect(hc.currentStatus.get('svc1')).toBe(latestRaw);
+  });
+
+  test('a long steady-green run produces exactly ONE emit (no per-probe spam)', () => {
+    for (let i = 0; i < 50; i++) hc.recordStatus('svc1', makeUp());
+    expect(emitSpy).toHaveBeenCalledTimes(1);
+  });
+
+  test('a long steady-green-then-steady-red transition: 1 emit (up), 1 emit (red)', () => {
+    for (let i = 0; i < 10; i++) hc.recordStatus('svc1', makeUp());
+    expect(emitSpy).toHaveBeenCalledTimes(1);
+    hc.recordStatus('svc1', makeDown());
+    hc.recordStatus('svc1', makeDown());      // flips to red
+    expect(emitSpy).toHaveBeenCalledTimes(2);
+    for (let i = 0; i < 10; i++) hc.recordStatus('svc1', makeDown());
+    expect(emitSpy).toHaveBeenCalledTimes(2); // no further broadcasts
+  });
+
+  test('DOWN_THRESHOLD env var is honored', () => {
+    process.env.HEALTH_DOWN_THRESHOLD = '3';
+    jest.resetModules();
+    const HC2Module = require('../src/monitoring/health-checker');
+    // Module is a singleton with DOWN_THRESHOLD captured at module load —
+    // resetModules gives us a fresh module-level instance with the new env.
+    const hc2 = HC2Module;
+    hc2.displayedStatus = new Map();
+    hc2.consecutiveSinceChange = new Map();
+    hc2.currentStatus = new Map();
+    hc2.history = {};
+    hc2.removeAllListeners('status-check');
+    const spy = jest.fn();
+    hc2.on('status-check', spy);
+    hc2.recordStatus('svc1', makeUp());
+    hc2.recordStatus('svc1', makeDown());      // 1
+    hc2.recordStatus('svc1', makeDown());      // 2 — still green (need 3)
+    expect(spy).toHaveBeenCalledTimes(1);
+    expect(hc2.displayedStatus.get('svc1').status).toBe('up');
+    hc2.recordStatus('svc1', makeDown());      // 3 — flip
+    expect(spy).toHaveBeenCalledTimes(2);
+    expect(hc2.displayedStatus.get('svc1').status).toBe('down');
+  });
+
+  test.each(['not-a-number', '0', '-2', '1.5'])('malformed DOWN_THRESHOLD %s falls back to 2', value => {
+    process.env.HEALTH_DOWN_THRESHOLD = value;
+    jest.resetModules();
+    const hc2 = require('../src/monitoring/health-checker');
+    hc2.displayedStatus = new Map();
+    hc2.consecutiveSinceChange = new Map();
+    hc2.currentStatus = new Map();
+    hc2.history = {};
+    hc2.removeAllListeners('status-check');
+    const spy = jest.fn();
+    hc2.on('status-check', spy);
+    hc2.recordStatus('svc1', makeUp());
+    hc2.recordStatus('svc1', makeDown());
+    expect(spy).toHaveBeenCalledTimes(1);
+    hc2.recordStatus('svc1', makeDown());
+    expect(spy).toHaveBeenCalledTimes(2);
+  });
+
+  test('UP_THRESHOLD env var greater than 1 is honored', () => {
+    process.env.HEALTH_UP_THRESHOLD = '2';
+    jest.resetModules();
+    const hc2 = require('../src/monitoring/health-checker');
+    hc2.displayedStatus = new Map();
+    hc2.consecutiveSinceChange = new Map();
+    hc2.currentStatus = new Map();
+    hc2.history = {};
+    hc2.removeAllListeners('status-check');
+    const spy = jest.fn();
+    hc2.on('status-check', spy);
+    hc2.recordStatus('svc1', makeDown());
+    hc2.recordStatus('svc1', makeUp());
+    expect(spy).toHaveBeenCalledTimes(1);
+    expect(hc2.displayedStatus.get('svc1').status).toBe('down');
+    hc2.recordStatus('svc1', makeUp());
+    expect(spy).toHaveBeenCalledTimes(2);
+    expect(hc2.displayedStatus.get('svc1').status).toBe('up');
+  });
+
+  test.each(['not-a-number', '0', '-2', '1.5'])('malformed UP_THRESHOLD %s falls back to 1', value => {
+    process.env.HEALTH_UP_THRESHOLD = value;
+    jest.resetModules();
+    const hc2 = require('../src/monitoring/health-checker');
+    hc2.displayedStatus = new Map();
+    hc2.consecutiveSinceChange = new Map();
+    hc2.currentStatus = new Map();
+    hc2.history = {};
+    hc2.removeAllListeners('status-check');
+    const spy = jest.fn();
+    hc2.on('status-check', spy);
+    hc2.recordStatus('svc1', makeDown());
+    hc2.recordStatus('svc1', makeUp());
+    expect(spy).toHaveBeenCalledTimes(2);
+    expect(hc2.displayedStatus.get('svc1').status).toBe('up');
+  });
+
+  test('removeService clears hysteresis state before the same ID is re-added', () => {
+    hc.config.services.svc1 = { name: 'Service 1' };
+    hc.recordStatus('svc1', makeUp());
+    hc.recordStatus('svc1', makeDown());
+    expect(hc.displayedStatus.has('svc1')).toBe(true);
+    expect(hc.consecutiveSinceChange.get('svc1')).toBe(1);
+    hc.consecutiveFailures.set('svc1', 3);
+    const timer = setTimeout(() => {}, 60_000);
+    hc.serviceTimers.set('svc1', timer);
+
+    hc.saveConfig = jest.fn();
+    hc.removeService('svc1');
+
+    expect(hc.displayedStatus.has('svc1')).toBe(false);
+    expect(hc.consecutiveSinceChange.has('svc1')).toBe(false);
+    expect(hc.currentStatus.has('svc1')).toBe(false);
+    expect(hc.consecutiveFailures.has('svc1')).toBe(false);
+    expect(hc.serviceTimers.has('svc1')).toBe(false);
+
+    hc.config.services.svc1 = { name: 'Service 1 re-added' };
+    const emitSpyAfterReAdd = jest.fn();
+    hc.on('status-check', emitSpyAfterReAdd);
+    hc.recordStatus('svc1', makeDown());
+
+    expect(emitSpyAfterReAdd).toHaveBeenCalledTimes(1);
+    expect(hc.displayedStatus.get('svc1').status).toBe('down');
+    expect(hc.consecutiveSinceChange.has('svc1')).toBe(false);
+  });
+});

dashcaddy-api/__tests__/health-checker-incident-parity.test.js

@@ -0,0 +1,186 @@
+/**
+ * Tests for DC-090: outage incidents follow the DISPLAYED (post-hysteresis)
+ * status — the same signal that flips the dashboard badge.
+ *
+ *   - A single raw "down" blip that hysteresis suppresses opens NO outage
+ *     incident (the DC-089-noted raw-transition bug).
+ *   - A suppressed blip does not resolve a real open outage (UP_THRESHOLD=2).
+ *   - DOWN_THRESHOLD consecutive downs open exactly ONE outage incident.
+ *   - The incident payload carries the displayed snapshot, not the raw probe.
+ *   - Direct callers without hysteresis state keep legacy raw semantics.
+ *
+ * The probe() helper replicates checkService's exact call order: capture the
+ * pre-probe raw + displayed state, recordStatus (updates both maps), then
+ * checkForIncidents with both previous states.
+ */
+
+'use strict';
+
+const path = require('path');
+const fs = require('fs');
+const os = require('os');
+
+// Use an isolated data dir so test history doesn't pollute the real one.
+const tmpDir = fs.mkdtempSync(path.join(os.tmpdir(), 'dashcaddy-incpar-'));
+process.env.HEALTH_DATA_DIR = tmpDir;
+process.env.HEALTH_CONFIG_FILE = path.join(tmpDir, 'health-config.json');
+process.env.HEALTH_HISTORY_FILE = path.join(tmpDir, 'health-history.json');
+
+// Module exports a singleton instance, not a class. Reset per-test state by
+// replacing the relevant maps on the singleton in beforeEach.
+const healthCheckerSingleton = require('../src/monitoring/health-checker');
+const originalDownThreshold = process.env.HEALTH_DOWN_THRESHOLD;
+const originalUpThreshold = process.env.HEALTH_UP_THRESHOLD;
+
+function restoreEnv(name, value) {
+  if (value === undefined) delete process.env[name];
+  else process.env[name] = value;
+}
+
+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-090: outage incidents follow the displayed (hysteresis) status', () => {
+  let hc;
+  let incidentCreatedSpy;
+  let incidentResolvedSpy;
+
+  beforeEach(() => {
+    healthCheckerSingleton.displayedStatus = new Map();
+    healthCheckerSingleton.consecutiveSinceChange = new Map();
+    healthCheckerSingleton.currentStatus = new Map();
+    healthCheckerSingleton.history = {};
+    healthCheckerSingleton.incidents = [];
+    healthCheckerSingleton.removeAllListeners('incident-created');
+    healthCheckerSingleton.removeAllListeners('incident-resolved');
+    incidentCreatedSpy = jest.fn();
+    incidentResolvedSpy = jest.fn();
+    healthCheckerSingleton.on('incident-created', incidentCreatedSpy);
+    healthCheckerSingleton.on('incident-resolved', incidentResolvedSpy);
+    hc = healthCheckerSingleton;
+  });
+
+  afterEach(() => {
+    restoreEnv('HEALTH_DOWN_THRESHOLD', originalDownThreshold);
+    restoreEnv('HEALTH_UP_THRESHOLD', originalUpThreshold);
+  });
+
+  afterAll(() => {
+    fs.rmSync(tmpDir, { recursive: true, force: true });
+  });
+
+  // Replicates checkService's record+incident sequence for one raw probe.
+  function probe(status, config = {}) {
+    const previousStatus = hc.currentStatus.get(status.serviceId);
+    const previousDisplayed = hc.displayedStatus.get(status.serviceId) || null;
+    hc.recordStatus(status.serviceId, status);
+    hc.checkForIncidents(status.serviceId, status, config, previousStatus, previousDisplayed);
+  }
+
+  test('a single down blip between two ups opens NO outage incident', () => {
+    probe(makeUp());      // baseline: displayed up
+    probe(makeDown());    // blip — hysteresis keeps displayed up
+    probe(makeUp());      // recovered
+    expect(hc.incidents).toHaveLength(0);
+    expect(incidentCreatedSpy).not.toHaveBeenCalled();
+  });
+
+  test('DOWN_THRESHOLD consecutive downs open exactly one outage incident (critical)', () => {
+    probe(makeUp());
+    probe(makeDown());    // counter=1, displayed still up
+    probe(makeDown());    // counter=2 → displayed flips down → incident
+    expect(hc.incidents).toHaveLength(1);
+    const incident = hc.incidents[0];
+    expect(incident.type).toBe('outage');
+    expect(incident.severity).toBe('critical');
+    expect(incident.status).toBe('open');
+    expect(incidentCreatedSpy).toHaveBeenCalledTimes(1);
+
+    probe(makeDown());    // still down — no new transition, no second incident
+    expect(hc.incidents).toHaveLength(1);
+    expect(incident.occurrences).toBe(1);  // occurrences count displayed flips, not raw probes
+    expect(incidentCreatedSpy).toHaveBeenCalledTimes(1);
+  });
+
+  test('the outage incident payload carries the displayed snapshot, not the raw blip', () => {
+    probe(makeUp());
+    const blip = makeDown();
+    blip.statusCode = 599;
+    probe(blip);          // suppressed blip — must not appear in any incident
+    probe(makeDown());    // flip
+    expect(hc.incidents).toHaveLength(1);
+    // The incident's details snapshot is the probe that FLIPPED the displayed
+    // state (the second down), not the earlier suppressed blip.
+    expect(hc.incidents[0].details.statusCode).not.toBe(599);
+  });
+
+  test('a suppressed up blip does not resolve a real open outage (UP_THRESHOLD=2)', () => {
+    process.env.HEALTH_UP_THRESHOLD = '2';
+    jest.resetModules();
+    const hc2 = require('../src/monitoring/health-checker');
+    hc2.displayedStatus = new Map();
+    hc2.consecutiveSinceChange = new Map();
+    hc2.currentStatus = new Map();
+    hc2.history = {};
+    hc2.incidents = [];
+    hc2.removeAllListeners('incident-created');
+    hc2.removeAllListeners('incident-resolved');
+
+    const p2 = (status) => {
+      const prevRaw = hc2.currentStatus.get(status.serviceId);
+      const prevDisp = hc2.displayedStatus.get(status.serviceId) || null;
+      hc2.recordStatus(status.serviceId, status);
+      hc2.checkForIncidents(status.serviceId, status, {}, prevRaw, prevDisp);
+    };
+
+    p2(makeUp());
+    p2(makeDown());
+    p2(makeDown());       // displayed down → outage opens
+    expect(hc2.incidents).toHaveLength(1);
+    expect(hc2.incidents[0].status).toBe('open');
+
+    p2(makeUp());         // counter=1 < UP_THRESHOLD=2 → displayed still down
+    expect(hc2.displayedStatus.get('svc1').status).toBe('down');
+    expect(hc2.incidents[0].status).toBe('open');   // NOT resolved by the blip
+
+    p2(makeUp());         // counter=2 → displayed up → incident resolves
+    expect(hc2.displayedStatus.get('svc1').status).toBe('up');
+    expect(hc2.incidents[0].status).toBe('resolved');
+  });
+
+  test('legacy direct callers (no displayed state) keep raw transition semantics', () => {
+    hc.currentStatus.set('svc1', { status: 'up' });
+    const status = { status: 'down', timestamp: new Date().toISOString(), responseTime: 100 };
+    hc.checkForIncidents('svc1', status, {});   // 4-arg call, no previousDisplayed
+    expect(hc.incidents).toHaveLength(1);
+    expect(hc.incidents[0].type).toBe('outage');
+  });
+
+  test('slow-response detection still fires per-probe regardless of hysteresis', () => {
+    const slowUp = makeUp();
+    slowUp.responseTime = 6000;
+    probe(slowUp, { slowResponseThreshold: 5000 });
+    expect(hc.incidents.some(i => i.type === 'slow-response')).toBe(true);
+  });
+});

dashcaddy-api/__tests__/health-checker.test.js

@@ -203,6 +203,55 @@ describe('HealthChecker', () => {
       expect(result.error).toBe('ECONNREFUSED');
     });
 
+    it('opens and resolves an outage incident across real checkService transitions', async () => {
+      // DC-090: incidents follow the DISPLAYED (post-hysteresis) status.
+      // DOWN_THRESHOLD defaults to 2, so it takes two consecutive failed
+      // probes to flip displayed down and open the outage; one up probe
+      // (UP_THRESHOLD=1) resolves it.
+      healthChecker._doRequest = jest.fn()
+        .mockResolvedValueOnce({ healthy: true, statusCode: 200, message: 'ok', details: {} })
+        .mockResolvedValueOnce({ healthy: false, statusCode: 500, message: 'down', details: {} })
+        .mockResolvedValueOnce({ healthy: false, statusCode: 500, message: 'down', details: {} })
+        .mockResolvedValueOnce({ healthy: true, statusCode: 200, message: 'ok', details: {} });
+
+      const config = { url: 'http://test.local' };
+      await healthChecker.checkService('svc1', config);
+      await healthChecker.checkService('svc1', config);
+      expect(healthChecker.incidents).toHaveLength(0);   // one down alone: suppressed blip
+
+      await healthChecker.checkService('svc1', config);  // second down flips displayed → open
+      expect(healthChecker.incidents).toHaveLength(1);
+      expect(healthChecker.incidents[0]).toMatchObject({
+        serviceId: 'svc1',
+        type: 'outage',
+        status: 'open'
+      });
+
+      await healthChecker.checkService('svc1', config);  // up resolves
+      expect(healthChecker.incidents[0].status).toBe('resolved');
+      expect(healthChecker.incidents[0].resolvedAt).toBeDefined();
+    });
+
+    it('does not resurrect state when an in-flight probe resolves after removal', async () => {
+      let resolveProbe;
+      healthChecker.config.services.svc1 = { url: 'http://test.local' };
+      healthChecker._doRequest = jest.fn(() => new Promise(resolve => {
+        resolveProbe = resolve;
+      }));
+
+      const pending = healthChecker.checkService('svc1', healthChecker.config.services.svc1);
+      healthChecker.saveConfig = jest.fn();
+      healthChecker.removeService('svc1');
+      resolveProbe({ healthy: true, statusCode: 200, message: 'late', details: {} });
+      await pending;
+
+      expect(healthChecker.currentStatus.has('svc1')).toBe(false);
+      expect(healthChecker.displayedStatus.has('svc1')).toBe(false);
+      expect(healthChecker.consecutiveFailures.has('svc1')).toBe(false);
+      expect(healthChecker.history.svc1).toBeUndefined();
+      expect(healthChecker.incidents).toEqual([]);
+    });
+
     it('increments consecutive failures on error', async () => {
       healthChecker._doRequest = jest.fn().mockRejectedValue(new Error('fail'));
 
@@ -549,6 +598,113 @@ describe('HealthChecker', () => {
     });
   });
 
+  describe('DC-088: removeService generation tombstones + incident closure', () => {
+    it('does not leak a serviceGenerations entry and records a tombstone', () => {
+      healthChecker.configureService('svc1', { url: 'http://test.local' });
+      expect(healthChecker.serviceGenerations.has('svc1')).toBe(true);
+
+      healthChecker.removeService('svc1');
+
+      expect(healthChecker.serviceGenerations.has('svc1')).toBe(false);
+      const tomb = healthChecker.removedGenerations.get('svc1');
+      expect(tomb).toBeDefined();
+      expect(tomb.generation).toBeGreaterThan(0);
+      expect(tomb.removedAt).toBeGreaterThan(0);
+    });
+
+    it('re-added service gets a strictly higher generation (no ABA)', () => {
+      healthChecker.configureService('svc1', { url: 'http://test.local' });
+      const gen1 = healthChecker.serviceGenerations.get('svc1');
+
+      healthChecker.removeService('svc1');
+      healthChecker.configureService('svc1', { url: 'http://test.local/v2' });
+
+      const gen2 = healthChecker.serviceGenerations.get('svc1');
+      expect(gen2).toBeGreaterThan(gen1);
+      expect(healthChecker.removedGenerations.has('svc1')).toBe(false);
+    });
+
+    it('closes open incidents for the removed service as resolved', () => {
+      healthChecker.saveConfig = jest.fn();
+      healthChecker.incidents.push({
+        id: 'incident-test-1',
+        serviceId: 'svc1',
+        type: 'outage',
+        status: 'open',
+        createdAt: new Date(Date.now() - 60_000).toISOString()
+      });
+      healthChecker.incidents.push({
+        id: 'incident-other',
+        serviceId: 'svc2',
+        type: 'outage',
+        status: 'open',
+        createdAt: new Date(Date.now() - 60_000).toISOString()
+      });
+      const resolvedSpy = jest.fn();
+      healthChecker.on('incident-resolved', resolvedSpy);
+
+      healthChecker.removeService('svc1');
+
+      const closed = healthChecker.incidents.find(i => i.id === 'incident-test-1');
+      expect(closed.status).toBe('resolved');
+      expect(closed.resolvedBy).toBe('service-removed');
+      expect(closed.resolvedAt).toBeDefined();
+      expect(closed.duration).toBeGreaterThan(0);
+      expect(healthChecker.incidents.find(i => i.id === 'incident-other').status).toBe('open');
+      expect(resolvedSpy).toHaveBeenCalledTimes(1);
+    });
+
+    it('in-flight probe captured before removal is discarded via tombstone', async () => {
+      let resolveProbe;
+      healthChecker.config.services.svc1 = { url: 'http://test.local' };
+      healthChecker._doRequest = jest.fn(() => new Promise(resolve => {
+        resolveProbe = resolve;
+      }));
+
+      const pending = healthChecker.checkService('svc1', healthChecker.config.services.svc1);
+      healthChecker.saveConfig = jest.fn();
+      healthChecker.removeService('svc1');
+      resolveProbe({ healthy: true, statusCode: 200, message: 'late', details: {} });
+      await pending;
+
+      expect(healthChecker.currentStatus.has('svc1')).toBe(false);
+      expect(healthChecker.consecutiveFailures.has('svc1')).toBe(false);
+    });
+
+    it('a rejected in-flight probe after removal does not re-create failure state', async () => {
+      let rejectProbe;
+      healthChecker.config.services.svc1 = { url: 'http://test.local' };
+      healthChecker._doRequest = jest.fn(() => new Promise((resolve, reject) => {
+        rejectProbe = reject;
+      }));
+
+      const pending = healthChecker.checkService('svc1', healthChecker.config.services.svc1);
+      healthChecker.saveConfig = jest.fn();
+      healthChecker.removeService('svc1');
+      rejectProbe(new Error('late failure'));
+      await pending;
+
+      expect(healthChecker.consecutiveFailures.has('svc1')).toBe(false);
+      expect(healthChecker.currentStatus.has('svc1')).toBe(false);
+    });
+
+    it('sweeps expired tombstones in cleanupHistory', () => {
+      healthChecker.removedGenerations.set('svc1', {
+        generation: 1,
+        removedAt: Date.now() - 60 * 60 * 1000 // 1h ago, TTL default 10m
+      });
+      healthChecker.removedGenerations.set('svc2', {
+        generation: 2,
+        removedAt: Date.now() // fresh
+      });
+
+      healthChecker.cleanupHistory();
+
+      expect(healthChecker.removedGenerations.has('svc1')).toBe(false);
+      expect(healthChecker.removedGenerations.has('svc2')).toBe(true);
+    });
+  });
+
   describe('cleanupHistory', () => {
     it('removes entries older than retention period', () => {
       const old = new Date(Date.now() - 35 * 24 * 60 * 60 * 1000).toISOString(); // 35 days ago

dashcaddy-api/__tests__/health-endpoints.test.js

@@ -26,6 +26,19 @@ jest.mock('dockerode', () => {
 function buildApp({ configOk = true, servicesOk = true, dockerOk = true, caddyOk = true } = {}) {
   process.env.MOCK_DOCKER_DOWN = dockerOk ? '0' : '1';
 
+  // DC-087 — mirror src/app.js faithfully: the caddy check goes through
+  // fetchT (which injects the Origin header Caddy's enforce_origin allowlist
+  // requires), and is MOCKED so the suite is hermetic — no live request to a
+  // real Caddy admin on :2019. The previous raw-`fetch` mirror sent an
+  // Origin-less probe to the LIVE admin whenever the full suite ran on the
+  // prod host (adversarial cron every 30 min): 12 journal 403 lines per run,
+  // ~700/day of `client is not allowed to access from origin ''` noise,
+  // plus a false checks.caddy.ok=false in the mirrored readiness payload.
+  const fetchT = jest.spyOn(require('../src/utils/http'), 'fetchT')
+    .mockImplementation(async () => (caddyOk
+      ? { ok: true, status: 200 }
+      : { ok: false, status: 403 }));
+
   const app = express();
   const config = {
     CONFIG_FILE: '/tmp/dc-test-config.json',
@@ -103,9 +116,13 @@ function buildApp({ configOk = true, servicesOk = true, dockerOk = true, caddyOk
       allOk = false;
     }
 
+    // DC-087 — mirror src/app.js exactly (fetchT, not raw fetch). fetchT is
+    // mocked at buildApp() scope, so this stays hermetic: no live probe to a
+    // real Caddy admin (the old raw-fetch mirror 403-spammed the prod journal
+    // every time the adversarial cron ran the full suite on this host).
     try {
       const caddyUrl = config.CADDY_ADMIN_URL || 'http://localhost:2019';
-      const response = await fetch(`${caddyUrl}/config/apps/http/servers/srv0/listen`, { signal: AbortSignal.timeout(10000) });
+      const response = await fetchT(`${caddyUrl}/config/apps/http/servers/srv0/listen`, {}, 10000);
       checks.caddy = { ok: response.ok, status: response.status };
       if (!response.ok) allOk = false;
     } catch (e) {

dashcaddy-api/__tests__/health-probe-aliases.test.js

@@ -33,9 +33,18 @@ jest.mock('dockerode', () => {
 
 // Mirror the canonical handler block from src/app.js — if this drifts from
 // the real handler, these tests will start failing and force a sync.
-function buildApp({ configOk = true, servicesOk = true, dockerOk = true } = {}) {
+function buildApp({ configOk = true, servicesOk = true, dockerOk = true, caddyOk = true } = {}) {
   process.env.MOCK_DOCKER_DOWN = dockerOk ? '0' : '1';
 
+  // DC-087 — mirror src/app.js: caddy check via fetchT (Origin-injecting),
+  // mocked here so the suite is hermetic. The old raw-fetch mirror probed the
+  // LIVE Caddy admin on :2019 whenever the full suite ran on the prod host
+  // (adversarial cron): Origin-less → 403 → 12 journal error lines per run.
+  const fetchT = jest.spyOn(require('../src/utils/http'), 'fetchT')
+    .mockImplementation(async () => (caddyOk
+      ? { ok: true, status: 200 }
+      : { ok: false, status: 403 }));
+
   const app = express();
   const config = {
     CONFIG_FILE: '/tmp/dc-test-config.json',
@@ -108,8 +117,10 @@ function buildApp({ configOk = true, servicesOk = true, dockerOk = true } = {})
       allOk = false;
     }
     try {
+      // DC-087 — mirror src/app.js exactly: fetchT (mocked above), not raw
+      // fetch. Hermetic: no live request to a real Caddy admin.
       const caddyUrl = config.CADDY_ADMIN_URL || 'http://localhost:2019';
-      const response = await fetch(`${caddyUrl}/config/apps/http/servers/srv0/listen`, { signal: AbortSignal.timeout(10000) });
+      const response = await fetchT(`${caddyUrl}/config/apps/http/servers/srv0/listen`, {}, 10000);
       checks.caddy = { ok: response.ok, status: response.status };
       if (!response.ok) allOk = false;
     } catch (e) {

dashcaddy-api/__tests__/utils-http-caddy-admin-origin.test.js

@@ -118,6 +118,67 @@ describe('Caddyfile + utils/http.js — Origin header construction (DC-051)', ()
     expect(offenders).toEqual([]);
   });
 
+  test('all :2019 call sites in TESTS use fetchT or a mocked fetchT (not raw fetch)', () => {
+    // DC-087 — the same rule, extended into __tests__. The api-code walk above
+    // skips __tests__, which let two mirrored health-handler test files keep a
+    // raw await-fetch caddy probe long after src/app.js moved to fetchT. On a
+    // host where the suite runs alongside a live Caddy admin (the prod box
+    // runs the full jest suite every 30 min via a cron adversarial check),
+    // that Origin-less raw fetch 403-spammed the Caddy journal (~700
+    // client-not-allowed error lines per day) while the tests still passed —
+    // checks.caddy.ok=false was silently accepted as sandbox noise. Mirrors
+    // MUST call fetchT (mocked at buildApp scope for hermeticity). A raw
+    // await-fetch at a Caddy-admin-URL call site in a test is an offender.
+    // NOTE: keep this comment free of backticks — stripComments pairs
+    // backtick spans across lines, and a stray pair shields real code from
+    // the comment stripper (this test self-flagged its first draft).
+    //
+    // Detection is deliberately FILE-LEVEL, not call-window: the historical
+    // drift kept the fetch call itself token-free (the URL came from a
+    // caddyUrl variable defined on a PREVIOUS line from CADDY_ADMIN_URL),
+    // so a call-window regex never fired. Any raw await-fetch in a file
+    // that also references the Caddy admin anywhere is an offender.
+    // Escape hatch for future tests that intentionally assert Origin-less
+    // 403 behavior against their own local listener: put the marker
+    // DC-087-ALLOW-RAW-FETCH in the file and it is skipped.
+    const testsRoot = path.join(__dirname);
+    const offenders = [];
+    const skipped = [];
+    function walk(dir) {
+      for (const entry of fs.readdirSync(dir, { withFileTypes: true })) {
+        if (entry.name === 'node_modules') continue;
+        const p = path.join(dir, entry.name);
+        if (entry.isDirectory()) walk(p);
+        else if (entry.name.endsWith('.js')) {
+          const rawText = fs.readFileSync(p, 'utf8');
+          // Escape hatch (checked on RAW text so a comment marker works —
+          // comments are stripped below): a file carrying the
+          // DC-087-ALLOW-RAW-FETCH marker declares it intentionally
+          // raw-fetches the Caddy admin (e.g. asserting Origin-less 403
+          // against its own local listener). The guard file itself is
+          // always scanned (never skipped) so the hatch can't be used to
+          // blind this very test.
+          if (p !== __filename && /DC-087-ALLOW-RAW-FETCH/.test(rawText)) {
+            skipped.push(p);
+            continue;
+          }
+          const text = stripComments(rawText);
+          const hasAdminToken = /:2019|adminUrl|admin_api_url|CADDY_ADMIN/.test(text);
+          const hasRawAwaitFetch = /await\s+fetch\(/.test(text);
+          if (hasAdminToken && hasRawAwaitFetch) {
+            offenders.push(`${p}: raw await-fetch in a file referencing the Caddy admin (mock fetchT instead; documented escape-hatch marker available for intentional 403 tests)`);
+          }
+        }
+      }
+    }
+    walk(testsRoot);
+    if (skipped.length) {
+      // Visibility for hatch use — shows up in jest output for reviewers.
+      console.info('[DC-087 guard] escape-hatch skipped:', skipped.join(', '));
+    }
+    expect(offenders).toEqual([]);
+  });
+
   test('readiness handler in src/app.js probes the exact URL the watcher needs', () => {
     const raw = fs.readFileSync(
       path.join(__dirname, '../src/app.js'),
@@ -127,9 +188,9 @@ describe('Caddyfile + utils/http.js — Origin header construction (DC-051)', ()
     expect(raw).toMatch(/\/config\/apps\/http\/servers\/srv0\/listen/);
     // Goes through fetchT, NOT bare fetch — that's how the Origin injection
     // takes effect. Look at the 800 chars BEFORE the probe URL on the same
-    // line / call site — the call must be `fetchT(...)`, not `await fetch(...)`.
-    // (We look backward because the URL sits inside the call's argument list,
-    // so the call site comes before the URL token.)
+    // line / call site — the call must be fetchT(...), never a raw await of
+    // the global fetch. (We look backward because the URL sits inside the
+    // call's argument list, so the call site comes before the URL token.)
     const idx = raw.indexOf('srv0/listen');
     const around = raw.substr(Math.max(0, idx - 400), 800);
     expect(around).toMatch(/fetchT\(/);

dashcaddy-api/routes/auth/admin.js

@@ -29,26 +29,10 @@ const platformPaths = require('../../platform-paths');
 const { createUserStore } = require('../../src/security/user-store');
 const { createInviteStore } = require('../../src/security/invite-store');
 const emailSender = require('../../src/auth/providers/email-sender');
+const AuthProvider = require('../../src/auth/providers/base');
 const { ValidationError, NotFoundError, ForbiddenError, PaymentRequiredError } = require('../../src/utilities/errors');
 const { ok, successMessage } = require('../../src/utils/responses');
 
-/**
- * Build the URL an invitee should click. Mirrors EmailMagicLinkProvider's
- * _resolvePublicUrl logic — kept duplicated (not extracted) because the two
- * callers have slightly different link paths and the duplication is smaller
- * than the abstraction would be.
- */
-function _buildInviteUrl(req, siteConfig, token) {
-  if (siteConfig && siteConfig.publicBaseUrl) {
-    return siteConfig.publicBaseUrl.replace(/\/+$/, '') +
-      '/api/v1/auth/invites/' + encodeURIComponent(token) + '/accept';
-  }
-  const proto = (req.headers && req.headers['x-forwarded-proto']) || (req.protocol || 'https');
-  const host = (req.headers && (req.headers['x-forwarded-host'] || req.headers.host))
-    || (siteConfig && siteConfig.dashboardHost) || 'localhost:3001';
-  return `${proto}://${host}/api/v1/auth/invites/${encodeURIComponent(token)}/accept`;
-}
-
 function _requireAdmin(req, _res, next) {
   if (!req.user || req.user.role !== 'admin') {
     return next(new ForbiddenError('Admin role required'));
@@ -240,11 +224,21 @@ module.exports = function({ asyncHandler, errorResponse, log, session, dataDir }
     });
     if (!issued.ok) throw new ValidationError(issued.reason, 'email');
 
-    let deliveredVia = 'none';
-    const maskedEmail = email.replace(/(^.).+(@.*$)/, '$1***$2');
-    if (sendEmail !== false) {
-      // Best-effort send. If SMTP isn't configured, log to error.log (dev path).
-      const acceptUrl = _buildInviteUrl(req, /* siteConfig */ req.app.locals && req.app.locals.siteConfig, issued.token);
+    // Build the accept URL once — used both for the response and for email delivery.
+    const baseUrl = (req.app.locals && req.app.locals.siteConfig && req.app.locals.siteConfig.publicBaseUrl
+      ? req.app.locals.siteConfig.publicBaseUrl.replace(/\/+$/, '')
+      : ((req.headers['x-forwarded-proto'] || req.protocol || 'https') + '://' +
+         (req.headers['x-forwarded-host'] || req.headers.host || 'localhost:3001')));
+    const acceptUrl = baseUrl + '/api/v1/auth/invites/' + encodeURIComponent(issued.token) + '/accept';
+
+    // DC-085: link-first delivery. Default = no email, just hand the link back.
+    // Operators opt INTO email by sending { sendEmail: true } (or the admin UI
+    // checks the "Send email" checkbox). When SMTP is unconfigured AND the
+    // operator did opt in, we surface the failure as `deliveredVia: 'failed'`
+    // but NEVER leak the raw token into the server log — the link is already
+    // in the response, so the operator has a UI-side fallback.
+    let deliveredVia = 'manual';
+    if (sendEmail === true) {
       const ttlHoursOut = Math.round(issued.ttlMs / (60 * 60 * 1000));
       const text = _buildEmailText({ acceptUrl, ttlHours: ttlHoursOut, role: issued.role });
       const html = _buildEmailHtml({ acceptUrl, ttlHours: ttlHoursOut, role: issued.role });
@@ -254,33 +248,37 @@ module.exports = function({ asyncHandler, errorResponse, log, session, dataDir }
           await emailSender.sendEmail(smtpConfig, issued.email, 'You\'re invited to DashCaddy', text, html);
           deliveredVia = 'email';
         } else {
-          // Dev fallback — log the raw link so operators can grab it.
-          log.warn && log.warn('auth-invite-dev',
-            '[DC-048-DEV-INVITE-LINK] email=' + issued.email +
-            ' role=' + issued.role + ' url=' + acceptUrl);
-          deliveredVia = 'dev-console';
+          // Operator asked for email but SMTP isn't configured. Surface the
+          // failure cleanly; the link is still in the response so the
+          // operator can share it manually. Do NOT log the raw URL — it
+          // would duplicate what's already in the response and pollute the
+          // server log on every unconfigured-install invite.
+          log.warn && log.warn('auth-invite-send',
+            'invite send skipped: SMTP not configured (operator opted in)',
+            { inviteId: issued.id, email: AuthProvider.maskEmail(issued.email) || '[unmaskable-email]' });
+          deliveredVia = 'failed';
         }
       } catch (sendErr) {
         log.warn && log.warn('auth-invite-send',
-          'invite send failed: ' + (sendErr.message || String(sendErr)));
+          'invite send failed: ' + (sendErr.message || String(sendErr)),
+          { inviteId: issued.id });
         deliveredVia = 'failed';
       }
-    } else {
-      deliveredVia = 'manual';
     }
 
+    const maskedEmail = email.replace(/(^.).+(@.*$)/, '$1***$2');
+    const ttlHoursOut = Math.round(issued.ttlMs / (60 * 60 * 1000));
+    const shareText =
+      'Join my DashCaddy as ' + issued.role + ' — ' + acceptUrl +
+      ' — expires in ' + ttlHoursOut + 'h.';
+
     return ok(res, {
       id: issued.id,
       email: issued.email,
       role: issued.role,
       expiresAt: issued.expiresAt,
-      // The raw token is returned ONCE so the admin UI can show/copy the
-      // link. It is also embedded in the email when sendEmail !== false.
-      acceptUrl: (req.app.locals && req.app.locals.siteConfig && req.app.locals.siteConfig.publicBaseUrl
-        ? req.app.locals.siteConfig.publicBaseUrl.replace(/\/+$/, '')
-        : ((req.headers['x-forwarded-proto'] || req.protocol || 'https') + '://' +
-           (req.headers['x-forwarded-host'] || req.headers.host || 'localhost:3001'))) +
-        '/api/v1/auth/invites/' + encodeURIComponent(issued.token) + '/accept',
+      acceptUrl,
+      shareText,
       deliveredVia,
       maskedEmail,
     });
@@ -372,7 +370,7 @@ module.exports = function({ asyncHandler, errorResponse, log, session, dataDir }
 
     log.info && log.info('auth', 'invite accepted, user created', {
       userId: userResult.user.id,
-      email: userResult.user.email,
+      email: AuthProvider.maskEmail(userResult.user.email) || '[unmaskable-email]',
       role: userResult.user.role,
       inviteId: invite.id,
     });

dashcaddy-api/src/monitoring/health-checker.js

@@ -32,6 +32,30 @@ const CHECK_INTERVAL = parseInt(process.env.HEALTH_CHECK_INTERVAL || '30000', 10
 const MAX_CHECK_INTERVAL = parseInt(process.env.HEALTH_CHECK_MAX_INTERVAL || '300000', 10); // 5 minutes max backoff
 const MAX_ENTRIES_PER_SERVICE = parseInt(process.env.HEALTH_MAX_ENTRIES || '500', 10); // Cap to prevent disk explosion
 const HISTORY_RETENTION_DAYS = parseInt(process.env.HEALTH_HISTORY_RETENTION || '30', 10);
+// DC-088: how long a removal tombstone outlives the removal itself. Only needs
+// to cover the max in-flight probe lifetime (timeout + scheduling headroom);
+// swept by cleanupHistory so removed services cannot accumulate map entries.
+const REMOVED_GENERATION_TTL_MS = parseInt(process.env.HEALTH_REMOVED_GEN_TTL || '600000', 10);
+
+// DC-086: hysteresis thresholds for badge display.
+// The raw probe result can flap on a single transient blip (Caddy reload,
+// container CPU steal, network hiccup, mid-flight TLS handshake). Showing
+// every probe result as-is to the dashboard creates the "perpetual flicker"
+// UX. Asymmetric thresholds: going red is slow (don't false-alarm), going
+// green is fast (don't keep showing red after recovery).
+//   - DOWN_THRESHOLD = N consecutive "down" probes before the badge flips to red
+//   - UP_THRESHOLD   = N consecutive "up" probes before the badge flips back to green
+// Single probe flips to green on purpose — false-positive-green is much less
+// painful than perpetual-red (operators notice red, ignore green).
+function readPositiveIntEnv(name, fallback) {
+  const raw = process.env[name];
+  if (raw === undefined || raw === '') return fallback;
+  const value = Number(raw);
+  return Number.isSafeInteger(value) && value >= 1 ? value : fallback;
+}
+
+const DOWN_THRESHOLD = readPositiveIntEnv('HEALTH_DOWN_THRESHOLD', 2);
+const UP_THRESHOLD = readPositiveIntEnv('HEALTH_UP_THRESHOLD', 1);
 
 class HealthChecker extends EventEmitter {
   constructor() {
@@ -39,11 +63,27 @@ class HealthChecker extends EventEmitter {
     this.config = this.loadConfig();
     this.history = this.loadHistory();
     this.currentStatus = new Map();
+    // DC-086: the status the dashboard SHOULD display (post-hysteresis).
+    // Distinct from currentStatus, which is the latest raw probe result.
+    this.displayedStatus = new Map();
+    // DC-086: counter of consecutive healthy/unhealthy probes since the
+    // last displayed-status change. Reset to 0 whenever displayed status flips.
+    this.consecutiveSinceChange = new Map();
     this.incidents = [];
     this.checking = false;
     this.checkInterval = null;
     this.consecutiveFailures = new Map(); // serviceId -> failure count
     this.serviceTimers = new Map(); // serviceId -> timer for per-service backoff
+    // Invalidate probe completions that race with removal/reconfiguration.
+    this.serviceGenerations = new Map(); // serviceId -> configuration generation
+    // DC-088: monotonically increasing sequence so generation numbers can never
+    // repeat across remove -> re-add cycles (prevents ABA on the stale check).
+    this.generationSeq = 0;
+    // DC-088: serviceId -> { generation, removedAt } tombstones. A live entry in
+    // serviceGenerations means the service is (re)configured; a tombstone with a
+    // HIGHER generation than the captured one marks the capture as stale. Entry
+    // is deleted when the service is removed, so the live map cannot leak.
+    this.removedGenerations = new Map();
   }
 
   /**
@@ -111,11 +151,27 @@ class HealthChecker extends EventEmitter {
     this.cleanupHistory();
   }
 
+  /**
+   * DC-088: true when a probe's captured generation no longer matches the
+   * service's current configuration state. A live serviceGenerations entry
+   * must match exactly. With no live entry the service was never configured
+   * in this process (disk-loaded / direct callers) — stale only if a removal
+   * tombstone with a HIGHER generation exists.
+   */
+  _isStaleCapture(serviceId, generation) {
+    if (this.serviceGenerations.has(serviceId)) {
+      return this.serviceGenerations.get(serviceId) !== generation;
+    }
+    const tomb = this.removedGenerations.get(serviceId);
+    return Boolean(tomb && tomb.generation > generation);
+  }
+
   /**
    * Check a single service
    */
   async checkService(serviceId, config) {
     const startTime = Date.now();
+    const generation = this.serviceGenerations.get(serviceId) || 0;
 
     try {
       const result = await this.performHealthCheck(config);
@@ -131,6 +187,10 @@ class HealthChecker extends EventEmitter {
         details: result.details
       };
 
+      if (this._isStaleCapture(serviceId, generation)) {
+        return status;
+      }
+
       // Track consecutive failures for exponential backoff
       if (result.healthy) {
         this.consecutiveFailures.delete(serviceId);
@@ -138,16 +198,15 @@ class HealthChecker extends EventEmitter {
         this.consecutiveFailures.set(serviceId, (this.consecutiveFailures.get(serviceId) || 0) + 1);
       }
 
+      const previousStatus = this.currentStatus.get(serviceId);
+      const previousDisplayed = this.displayedStatus.get(serviceId);
       this.recordStatus(serviceId, status);
-      this.checkForIncidents(serviceId, status, config);
+      this.checkForIncidents(serviceId, status, config, previousStatus, previousDisplayed);
 
       return status;
     } catch (error) {
       const responseTime = Date.now() - startTime;
 
-      // Increment failure count for backoff
-      this.consecutiveFailures.set(serviceId, (this.consecutiveFailures.get(serviceId) || 0) + 1);
-
       const status = {
         serviceId,
         timestamp: new Date().toISOString(),
@@ -156,8 +215,18 @@ class HealthChecker extends EventEmitter {
         error: error.message
       };
 
+      if (this._isStaleCapture(serviceId, generation)) {
+        return status;
+      }
+
+      // Increment failure count for backoff — only after the result is known
+      // to be non-stale, so a removed service cannot re-create map entries.
+      this.consecutiveFailures.set(serviceId, (this.consecutiveFailures.get(serviceId) || 0) + 1);
+
+      const previousStatus = this.currentStatus.get(serviceId);
+      const previousDisplayed = this.displayedStatus.get(serviceId);
       this.recordStatus(serviceId, status);
-      this.checkForIncidents(serviceId, status, config);
+      this.checkForIncidents(serviceId, status, config, previousStatus, previousDisplayed);
 
       return status;
     }
@@ -273,27 +342,109 @@ class HealthChecker extends EventEmitter {
     return true;
   }
 
+  /**
+   * Compute the displayed status for a service given the latest raw probe
+   * result. Applies asymmetric hysteresis:
+   *   - Going DOWN: requires DOWN_THRESHOLD (default 2) consecutive "down"
+   *     probes since the last display-state change. A single blip keeps the
+   *     badge green.
+   *   - Going UP: requires UP_THRESHOLD (default 1) consecutive "up" probes.
+   *     Any single "up" after a down streak flips back to green so the badge
+   *     doesn't linger red after the service has recovered.
+   *
+   * Returns the displayed status object (same shape as the raw status) so
+   * recordStatus can use it both for the displayed map and as the broadcast
+   * payload when the displayed status actually changes.
+   */
+  _computeDisplayedStatus(serviceId, rawStatus) {
+    const currentDisplayed = this.displayedStatus.get(serviceId);
+    const previousStatus = currentDisplayed ? currentDisplayed.status : null;
+
+    // If no prior state, accept the raw probe as-is (first-check bootstrap).
+    if (!previousStatus) {
+      return rawStatus;
+    }
+
+    // Probe agrees with current displayed → no change, reset the counter so
+    // a brief blip doesn't accumulate against the displayed state.
+    if (rawStatus.status === previousStatus) {
+      this.consecutiveSinceChange.set(serviceId, 0);
+      return rawStatus;
+    }
+
+    // Probe disagrees with displayed. Bump the streak counter — this counts
+    // CONSECUTIVE probes that disagree with what's shown, regardless of
+    // whether the raw value itself changed between probes. That's what
+    // makes "down, down" flip after threshold but "down, up, down" not flip.
+    const prev = this.consecutiveSinceChange.get(serviceId) || 0;
+    const next = prev + 1;
+
+    if (rawStatus.status === 'down') {
+      // Going DOWN: need DOWN_THRESHOLD consecutive probes that disagree
+      // with the displayed "up" state.
+      if (previousStatus === 'up' && next < DOWN_THRESHOLD) {
+        this.consecutiveSinceChange.set(serviceId, next);
+        // Keep the last internally-consistent displayed snapshot. Mixing the
+        // raw failure metadata with status="up" would expose contradictory
+        // API data (for example statusCode=500 on an "up" service).
+        return currentDisplayed;
+      }
+      // Threshold met (or already down) — flip to red.
+      this.consecutiveSinceChange.set(serviceId, 0);
+      return rawStatus;
+    }
+
+    // rawStatus.status === 'up' (must be — the equal-to-displayed case above
+    // already returned). Going UP after a down streak: need UP_THRESHOLD.
+    if (previousStatus === 'down' && next < UP_THRESHOLD) {
+      this.consecutiveSinceChange.set(serviceId, next);
+      return currentDisplayed;
+    }
+    this.consecutiveSinceChange.set(serviceId, 0);
+    return rawStatus;
+  }
+
   /**
    * Record service status
+   *
+   * DC-086: history + consecutiveFailures are updated for EVERY probe
+   * (operators want full probe history for postmortems). The dashboard's
+   * `status-check` event is only emitted when the DISPLAYED status changes,
+   * so the badge stops re-rendering on every probe.
    */
   recordStatus(serviceId, status) {
-    // Update current status
+    // Update current (raw) status — used by checkForIncidents and history.
     this.currentStatus.set(serviceId, status);
 
-    // Add to history
+    // Add raw probe to history (full fidelity — operators rely on this).
     if (!this.history[serviceId]) {
       this.history[serviceId] = [];
     }
 
     this.history[serviceId].push(status);
-    
+
     // Cap entries to prevent unbounded growth (disk explosion fix)
     if (this.history[serviceId].length > MAX_ENTRIES_PER_SERVICE) {
       this.history[serviceId] = this.history[serviceId].slice(-MAX_ENTRIES_PER_SERVICE);
     }
 
-    // Emit status event
-    this.emit('status-check', status);
+    // Compute the post-hysteresis displayed status; only emit when it changes.
+    // _computeDisplayedStatus compares the raw probe against the DISPLAYED
+    // status (not the previous raw status), so the "consecutive since
+    // change" counter doesn't depend on the order of writes here.
+    const displayed = this._computeDisplayedStatus(serviceId, status);
+    const previousDisplayed = this.displayedStatus.get(serviceId);
+    const displayChanged =
+      !previousDisplayed || previousDisplayed.status !== displayed.status;
+
+    this.displayedStatus.set(serviceId, displayed);
+
+    if (displayChanged) {
+      // Emit with the displayed status so the dashboard renders the same
+      // state the hysteresis just decided. The raw probe result is still
+      // in `history` and `currentStatus` for anyone who wants it.
+      this.emit('status-check', displayed);
+    }
 
     // Save history periodically
     if (Math.random() < 0.05) { // 5% chance (every ~20 checks)
@@ -304,11 +455,27 @@ class HealthChecker extends EventEmitter {
   /**
    * Check for incidents (downtime, slow response, etc.)
    */
-  checkForIncidents(serviceId, status, config) {
-    const previous = this.currentStatus.get(serviceId);
+  checkForIncidents(serviceId, status, config, previous = this.currentStatus.get(serviceId), previousDisplayed = null) {
     
-    // Check for status change (up -> down or down -> up)
-    if (previous && previous.status !== status.status) {
+    // DC-090: outage incidents follow the DISPLAYED (post-hysteresis) status —
+    // the same signal that flips the dashboard badge. A single raw "down"
+    // blip that hysteresis suppresses must not open a critical outage
+    // incident (and a suppressed blip must not resolve a real one). When the
+    // caller supplies the pre-probe displayed state (checkService always
+    // does), transitions are evaluated displayed-vs-displayed using the
+    // post-recordStatus state in this.displayedStatus. Direct callers with
+    // no hysteresis state (previousDisplayed === null) keep the legacy
+    // raw-probe transition semantics.
+    if (previousDisplayed) {
+      const displayed = this.displayedStatus.get(serviceId);
+      if (displayed && displayed.status !== previousDisplayed.status) {
+        if (displayed.status === 'down') {
+          this.createIncident(serviceId, 'outage', 'Service is down', displayed);
+        } else if (displayed.status === 'up') {
+          this.resolveIncident(serviceId, 'outage', displayed);
+        }
+      }
+    } else if (previous && previous.status !== status.status) {
       if (status.status === 'down') {
         this.createIncident(serviceId, 'outage', 'Service is down', status);
       } else if (status.status === 'up') {
@@ -445,19 +612,29 @@ class HealthChecker extends EventEmitter {
   }
 
   /**
-   * Get current status for all services
+   * Get current status for all services.
+   *
+   * DC-086: returns the DISPLAYED status (post-hysteresis), not the latest
+   * raw probe. A page reload should show the same badge state the live
+   * SSE stream is currently showing — otherwise an operator who reloads
+   * the page after a single blip sees red even though the hysteresis kept
+   * the badge green for them.
    */
   getCurrentStatus() {
     const result = {};
-    
-    for (const [serviceId, status] of this.currentStatus.entries()) {
+
+    for (const [serviceId, rawStatus] of this.currentStatus.entries()) {
       const config = this.config.services[serviceId];
       const uptime24h = this.calculateUptime(serviceId, 24);
       const uptime7d = this.calculateUptime(serviceId, 168);
       const avgResponseTime = this.calculateAverageResponseTime(serviceId, 24);
-      
+
+      // Prefer the displayed status if we've already computed one; fall back
+      // to the raw probe on the very first call (before recordStatus has run).
+      const displayed = this.displayedStatus.get(serviceId) || rawStatus;
+
       result[serviceId] = {
-        ...status,
+        ...displayed,
         name: config?.name || serviceId,
         uptime: {
           '24h': uptime24h,
@@ -467,7 +644,7 @@ class HealthChecker extends EventEmitter {
         sla: config?.sla
       };
     }
-    
+
     return result;
   }
 
@@ -530,6 +707,13 @@ class HealthChecker extends EventEmitter {
       this.config.services = {};
     }
 
+    // DC-088: monotonic instance-wide sequence — a re-added service can never
+    // recycle a previous generation number, and any older in-flight capture is
+    // invalidated by definition.
+    this.generationSeq += 1;
+    this.serviceGenerations.set(serviceId, this.generationSeq);
+    // Re-configuration supersedes any prior removal tombstone.
+    this.removedGenerations.delete(serviceId);
     this.config.services[serviceId] = {
       enabled: config.enabled !== false,
       name: config.name || serviceId,
@@ -552,12 +736,42 @@ class HealthChecker extends EventEmitter {
    * Remove service configuration
    */
   removeService(serviceId) {
+    // DC-088: tombstone the captured generation instead of leaking an entry.
+    // The live map entry is deleted; an in-flight probe captured BEFORE this
+    // point sees no live entry but a higher tombstone generation, so it is
+    // discarded. configureService clears the tombstone on re-add.
+    this.generationSeq += 1;
+    this.serviceGenerations.delete(serviceId);
+    this.removedGenerations.set(serviceId, {
+      generation: this.generationSeq,
+      removedAt: Date.now()
+    });
     if (this.config.services) {
       delete this.config.services[serviceId];
       this.saveConfig();
     }
 
+    // DC-088: open incidents for a removed service must not linger forever.
+    // Close them through the same resolve path a recovery would, annotated so
+    // history shows why (dashboard renders resolved incidents green + duration).
+    for (const incident of this.incidents) {
+      if (incident.serviceId === serviceId && incident.status === 'open') {
+        incident.status = 'resolved';
+        incident.resolvedAt = new Date().toISOString();
+        incident.duration = new Date(incident.resolvedAt) - new Date(incident.createdAt);
+        incident.resolvedBy = 'service-removed';
+        this.emit('incident-resolved', incident);
+        this.emit('log', 'info', `Incident closed by service removal: ${incident.id}`);
+      }
+    }
+
     this.currentStatus.delete(serviceId);
+    this.displayedStatus.delete(serviceId);
+    this.consecutiveSinceChange.delete(serviceId);
+    this.consecutiveFailures.delete(serviceId);
+    const timer = this.serviceTimers.get(serviceId);
+    if (timer) clearTimeout(timer);
+    this.serviceTimers.delete(serviceId);
     delete this.history[serviceId];
   }
 
@@ -576,6 +790,18 @@ class HealthChecker extends EventEmitter {
         this.history[serviceId] = this.history[serviceId].slice(-MAX_ENTRIES_PER_SERVICE);
       }
     }
+
+    // DC-088: sweep expired removal tombstones. After the TTL no probe that
+    // captured a pre-removal generation can still be in flight (timeout is
+    // bounded by performHealthCheck), so the tombstone has done its job.
+    if (this.removedGenerations.size > 0) {
+      const now = Date.now();
+      for (const [serviceId, tomb] of this.removedGenerations) {
+        if (now - tomb.removedAt > REMOVED_GENERATION_TTL_MS) {
+          this.removedGenerations.delete(serviceId);
+        }
+      }
+    }
   }
 
   /**

status/js/admin.js

@@ -216,8 +216,8 @@
       _el('input', { name: 'ttlHours', type: 'number', min: '1', max: '168', value: '24', style: 'padding:6px;width:80px' }),
     ));
     form.appendChild(_el('label', { style: 'display:flex;gap:4px;align-items:center;font-size:0.85rem' },
-      _el('input', { name: 'sendEmail', type: 'checkbox', checked: true }),
-      _el('span', { text: 'Send email' }),
+      _el('input', { name: 'sendEmail', type: 'checkbox', checked: false }),
+      _el('span', { text: 'Also send via email (optional)' }),
     ));
     form.appendChild(_el('button', { type: 'submit', class: 'btn-sm', style: 'padding:6px 12px', text: 'Issue invite' }));
     container.appendChild(form);
@@ -297,16 +297,77 @@
       },
     });
     banner.appendChild(copyBtn);
-    if (invite.deliveredVia === 'dev-console') {
+
+    // DC-085: pre-formatted message for one-tap paste into iMessage / WhatsApp /
+    // Telegram / SMS / Signal / Discord / paste-into-email. The operator can
+    // copy this as a sentence instead of dealing with the raw URL.
+    if (invite.shareText) {
+      const shareBlock = _el('div', { style: 'margin-top:12px' });
+      shareBlock.appendChild(_el('div', {
+        style: 'font-size:0.8rem;color:#86efac;margin-bottom:4px',
+        text: 'Share this message:',
+      }));
+      shareBlock.appendChild(_el('div', {
+        style: 'padding:8px;background:#000;border-radius:4px;color:#d1fae5;white-space:pre-wrap',
+        text: invite.shareText,
+      }));
+      const shareActions = _el('div', { style: 'margin-top:6px;display:flex;gap:6px;flex-wrap:wrap' });
+      const copyTextBtn = _el('button', {
+        class: 'btn-sm', style: 'padding:4px 10px',
+        text: 'Copy message',
+        onclick: async () => {
+          try {
+            await navigator.clipboard.writeText(invite.shareText);
+            copyTextBtn.textContent = 'Copied!';
+            setTimeout(() => { copyTextBtn.textContent = 'Copy message'; }, 2000);
+          } catch (e) {
+            window.errorHandler && window.errorHandler.show('Clipboard blocked: select the text manually.');
+          }
+        },
+      });
+      shareActions.appendChild(copyTextBtn);
+      // Native share sheet on mobile / supported browsers. Falls back silently
+      // (the copy buttons cover the same intent).
+      if (typeof navigator !== 'undefined' && typeof navigator.share === 'function') {
+        const nativeShareBtn = _el('button', {
+          class: 'btn-sm', style: 'padding:4px 10px',
+          text: 'Share via…',
+          onclick: async () => {
+            try {
+              await navigator.share({
+                title: 'DashCaddy invite',
+                text: invite.shareText,
+                url: invite.acceptUrl,
+              });
+            } catch (e) {
+              // User-cancelled throws AbortError — that's fine, just stay quiet.
+              if (e && e.name && e.name !== 'AbortError') {
+                window.errorHandler && window.errorHandler.show('Share failed: ' + e.message);
+              }
+            }
+          },
+        });
+        shareActions.appendChild(nativeShareBtn);
+      }
+      shareBlock.appendChild(shareActions);
+      banner.appendChild(shareBlock);
+    }
+
+    if (invite.deliveredVia === 'failed') {
       banner.appendChild(_el('p', {
         style: 'margin-top:8px;color:#fbbf24;font-size:0.8rem',
-        text: 'SMTP not configured — the invite was logged to the server console (search for [DC-048-DEV-INVITE-LINK]).',
+        text: 'Email could not be sent (SMTP not configured). Share the link above instead — it works the same way.',
       }));
     } else if (invite.deliveredVia === 'email') {
       banner.appendChild(_el('p', {
         style: 'margin-top:8px;color:#86efac;font-size:0.8rem',
         text: 'Email sent to ' + invite.email + '.',
       }));
+    } else if (invite.deliveredVia === 'manual') {
+      banner.appendChild(_el('p', {
+        style: 'margin-top:8px;color:#86efac;font-size:0.8rem',
+        text: 'Share the link above via text, chat, or any messenger.',
+      }));
     }
     parent.appendChild(banner);
   }

status/sw.js

@@ -1,4 +1,4 @@
-const CACHE = 'dashcaddy-shell-c25cea8485';
+const CACHE = 'dashcaddy-shell-497e1f671c';
 const PRECACHE = [
   '/',
   '/index.html',