Merge pull request 'feat: incident management and maintenance windows' (#17) from feat/incident-management into main

Reviewed-on: lerko/uptime#17

internal/metrics/prometheus.go

@@ -55,6 +55,15 @@ func Handler(eng *monitor.Engine) http.HandlerFunc {
 			writeGauge(&b, "upkeep_monitor_paused", labels(s), float64(val))
 		}
 
+		writeHelp(&b, "upkeep_monitor_maintenance", "gauge", "Whether the monitor is in a maintenance window (1) or not (0).")
+		for _, s := range sites {
+			val := 0
+			if eng.GetDisplayStatus(s) == "MAINT" {
+				val = 1
+			}
+			writeGauge(&b, "upkeep_monitor_maintenance", labels(s), float64(val))
+		}
+
 		writeHelp(&b, "upkeep_monitor_cert_expiry_timestamp_seconds", "gauge", "Unix timestamp when the SSL certificate expires.")
 		for _, s := range sites {
 			if !s.HasSSL || s.CertExpiry.IsZero() {

internal/metrics/prometheus_test.go

@@ -52,6 +52,16 @@ func (m *mockStore) UpdateNodeLastSeen(string) error                  { return n
 func (m *mockStore) DeleteNode(string) error                          { return nil }
 func (m *mockStore) SaveLog(string) error                             { return nil }
 func (m *mockStore) LoadLogs(int) ([]string, error)                   { return nil, nil }
+func (m *mockStore) GetActiveMaintenanceWindows() ([]models.MaintenanceWindow, error) {
+	return nil, nil
+}
+func (m *mockStore) GetAllMaintenanceWindows(int) ([]models.MaintenanceWindow, error) {
+	return nil, nil
+}
+func (m *mockStore) AddMaintenanceWindow(models.MaintenanceWindow) error { return nil }
+func (m *mockStore) EndMaintenanceWindow(int) error                      { return nil }
+func (m *mockStore) DeleteMaintenanceWindow(int) error                   { return nil }
+func (m *mockStore) IsMonitorInMaintenance(int) (bool, error)            { return false, nil }
 
 func TestMetricsHandler(t *testing.T) {
 	ms := &mockStore{

internal/models/models.go

@@ -67,8 +67,21 @@ type ProbeNode struct {
 	Version  string
 }
 
-type Backup struct {
-	Sites  []Site        `json:"sites"`
-	Alerts []AlertConfig `json:"alerts"`
-	Users  []User        `json:"users"`
+type MaintenanceWindow struct {
+	ID          int
+	MonitorID   int
+	Title       string
+	Description string
+	Type        string // "maintenance" or "incident"
+	StartTime   time.Time
+	EndTime     time.Time // zero = ongoing
+	CreatedBy   string
+	CreatedAt   time.Time
+}
+
+type Backup struct {
+	Sites              []Site              `json:"sites"`
+	Alerts             []AlertConfig       `json:"alerts"`
+	Users              []User              `json:"users"`
+	MaintenanceWindows []MaintenanceWindow `json:"maintenance_windows,omitempty"`
 }

internal/monitor/monitor.go

@@ -385,10 +385,16 @@ func (e *Engine) handleStatusChange(site models.Site, rawStatus string, code int
 		newState.FailureCount = site.MaxRetries + 1
 	}
 
+	inMaint := e.isInMaintenance(site.ID)
+
 	if site.Type == "http" && site.CheckSSL && site.HasSSL {
 		daysLeft := int(time.Until(site.CertExpiry).Hours() / 24)
 		if daysLeft <= site.ExpiryThreshold && !site.SentSSLWarning && rawStatus != "SSL EXP" {
-			e.triggerAlert(site.AlertID, "SSL WARNING", fmt.Sprintf("SSL for '%s' expires in %d days", site.Name, daysLeft))
+			if !inMaint {
+				e.triggerAlert(site.AlertID, "SSL WARNING", fmt.Sprintf("SSL for '%s' expires in %d days", site.Name, daysLeft))
+			} else {
+				e.AddLog(fmt.Sprintf("SSL warning for '%s' suppressed (maintenance)", site.Name))
+			}
 			newState.SentSSLWarning = true
 		} else if daysLeft > site.ExpiryThreshold {
 			newState.SentSSLWarning = false
@@ -405,14 +411,22 @@ func (e *Engine) handleStatusChange(site models.Site, rawStatus string, code int
 
 	isBroken := func(s string) bool { return s == "DOWN" || s == "SSL EXP" }
 	if !isBroken(site.Status) && isBroken(newState.Status) && newState.Status != "PENDING" {
-		msg := fmt.Sprintf("Monitor '%s' is DOWN (%s)", site.Name, rawStatus)
-		if site.Type == "push" {
-			msg = fmt.Sprintf("Push Monitor '%s' missed heartbeat.", site.Name)
+		if inMaint {
+			e.AddLog(fmt.Sprintf("Monitor '%s' is DOWN (alerts suppressed — maintenance)", site.Name))
+		} else {
+			msg := fmt.Sprintf("Monitor '%s' is DOWN (%s)", site.Name, rawStatus)
+			if site.Type == "push" {
+				msg = fmt.Sprintf("Push Monitor '%s' missed heartbeat.", site.Name)
+			}
+			e.triggerAlert(site.AlertID, "🚨 ALERT", msg)
 		}
-		e.triggerAlert(site.AlertID, "🚨 ALERT", msg)
 	}
 	if isBroken(site.Status) && newState.Status == "UP" {
-		e.triggerAlert(site.AlertID, "✅ RECOVERY", fmt.Sprintf("Monitor '%s' is UP", site.Name))
+		if !inMaint {
+			e.triggerAlert(site.AlertID, "✅ RECOVERY", fmt.Sprintf("Monitor '%s' is UP", site.Name))
+		} else {
+			e.AddLog(fmt.Sprintf("Monitor '%s' recovered (maintenance active, alert suppressed)", site.Name))
+		}
 	}
 }
 
@@ -432,6 +446,24 @@ func (e *Engine) triggerAlert(alertID int, title, message string) {
 	}
 }
 
+func (e *Engine) isInMaintenance(monitorID int) bool {
+	inMaint, err := e.db.IsMonitorInMaintenance(monitorID)
+	if err != nil {
+		return false
+	}
+	return inMaint
+}
+
+func (e *Engine) GetDisplayStatus(site models.Site) string {
+	if site.Paused {
+		return "PAUSED"
+	}
+	if e.isInMaintenance(site.ID) {
+		return "MAINT"
+	}
+	return site.Status
+}
+
 func (e *Engine) checkGroup(site models.Site) {
 	e.mu.RLock()
 	status := "UP"
@@ -445,7 +477,7 @@ func (e *Engine) checkGroup(site models.Site) {
 		if !child.Paused {
 			allPaused = false
 		}
-		if child.Paused {
+		if child.Paused || e.isInMaintenance(child.ID) {
 			continue
 		}
 		if child.Status == "DOWN" || child.Status == "SSL EXP" {

internal/server/server.go

@@ -35,6 +35,7 @@ var statusTpl = template.Must(template.New("status").Parse(`
 		.PENDING { background: #e0af68; color: #1a1b26; }
 		.SSL-EXP { background: #e0af68; color: #1a1b26; }
 		.PAUSED { background: #565f89; color: #c0caf5; }
+		.MAINT { background: #bb9af7; color: #1a1b26; }
 		.summary { display: flex; justify-content: center; gap: 16px; margin-bottom: 24px; font-size: 0.95em; font-weight: 600; }
 		.summary span { padding: 4px 12px; border-radius: 6px; }
 		.summary .s-up { color: #9ece6a; }
@@ -68,15 +69,17 @@ var statusTpl = template.Must(template.New("status").Parse(`
 		}
 
 		function renderSummary(sites) {
-			var up = 0, down = 0, paused = 0, total = sites.length;
+			var up = 0, down = 0, paused = 0, maint = 0, total = sites.length;
 			for (var i = 0; i < sites.length; i++) {
 				if (sites[i].Paused) { paused++; continue; }
+				if (sites[i].Status === 'MAINT') { maint++; continue; }
 				if (sites[i].Status === 'UP') up++;
 				else if (sites[i].Status === 'DOWN') down++;
 			}
 			var el = document.getElementById('summary');
 			var parts = ['<span class="s-total">' + up + '/' + total + ' UP</span>'];
 			if (down > 0) parts.push('<span class="s-down">' + down + ' DOWN</span>');
+			if (maint > 0) parts.push('<span style="color:#bb9af7">' + maint + ' MAINT</span>');
 			if (paused > 0) parts.push('<span class="s-paused">' + paused + ' PAUSED</span>');
 			el.innerHTML = parts.join('<span style="color:#383838">·</span>');
 		}
@@ -110,7 +113,7 @@ var statusTpl = template.Must(template.New("status").Parse(`
 			renderSummary(sites);
 			for (var i = 0; i < sites.length; i++) {
 				var s = sites[i];
-				var st = s.Paused ? 'PAUSED' : s.Status;
+				var st = s.Status === 'MAINT' ? 'MAINT' : s.Paused ? 'PAUSED' : s.Status;
 				var cls = cssClass(st);
 				var meta = esc(s.Type) + ' | ' + (s.Type === 'http' ? esc(s.URL) : 'Heartbeat Monitor');
 				var lc = s.LastCheck ? new Date(s.LastCheck).toLocaleTimeString('en-GB', {hour12: false}) : '—';
@@ -359,8 +362,24 @@ func Start(cfg ServerConfig, s store.Store, eng *monitor.Engine) {
 		mux.HandleFunc("/status", func(w http.ResponseWriter, r *http.Request) { renderStatusPage(w, cfg.Title, eng) })
 		mux.HandleFunc("/status/json", func(w http.ResponseWriter, r *http.Request) {
 			state := eng.GetLiveState()
+			activeWindows, _ := s.GetActiveMaintenanceWindows()
+			maintSet := make(map[int]bool)
+			allInMaint := false
+			for _, mw := range activeWindows {
+				if mw.Type != "maintenance" {
+					continue
+				}
+				if mw.MonitorID == 0 {
+					allInMaint = true
+				} else {
+					maintSet[mw.MonitorID] = true
+				}
+			}
 			for id, site := range state {
 				site.Token = ""
+				if allInMaint || maintSet[site.ID] || (site.ParentID > 0 && maintSet[site.ParentID]) {
+					site.Status = "MAINT"
+				}
 				state[id] = site
 			}
 			w.Header().Set("Content-Type", "application/json")

internal/store/postgres.go

@@ -56,6 +56,17 @@ func (d *PostgresDialect) CreateTablesSQL() []string {
 			message TEXT NOT NULL,
 			created_at TIMESTAMP DEFAULT NOW()
 		)`,
+		`CREATE TABLE IF NOT EXISTS maintenance_windows (
+			id SERIAL PRIMARY KEY,
+			monitor_id INTEGER DEFAULT 0,
+			title TEXT NOT NULL,
+			description TEXT DEFAULT '',
+			type TEXT DEFAULT 'maintenance',
+			start_time TIMESTAMP NOT NULL,
+			end_time TIMESTAMP,
+			created_by TEXT DEFAULT '',
+			created_at TIMESTAMP DEFAULT NOW()
+		)`,
 	}
 }
 
@@ -87,10 +98,12 @@ func (d *PostgresDialect) ImportWipe(tx *sql.Tx) {
 	tx.Exec("TRUNCATE TABLE sites RESTART IDENTITY CASCADE")
 	tx.Exec("TRUNCATE TABLE alerts RESTART IDENTITY CASCADE")
 	tx.Exec("TRUNCATE TABLE users RESTART IDENTITY CASCADE")
+	tx.Exec("TRUNCATE TABLE maintenance_windows RESTART IDENTITY CASCADE")
 }
 
 func (d *PostgresDialect) ImportResetSequences(tx *sql.Tx) {
 	tx.Exec("SELECT setval('sites_id_seq', (SELECT COALESCE(MAX(id), 1) FROM sites))")
 	tx.Exec("SELECT setval('alerts_id_seq', (SELECT COALESCE(MAX(id), 1) FROM alerts))")
 	tx.Exec("SELECT setval('users_id_seq', (SELECT COALESCE(MAX(id), 1) FROM users))")
+	tx.Exec("SELECT setval('maintenance_windows_id_seq', (SELECT COALESCE(MAX(id), 1) FROM maintenance_windows))")
 }

internal/store/sqlite.go

@@ -56,6 +56,17 @@ func (d *SQLiteDialect) CreateTablesSQL() []string {
 			message TEXT NOT NULL,
 			created_at DATETIME DEFAULT CURRENT_TIMESTAMP
 		)`,
+		`CREATE TABLE IF NOT EXISTS maintenance_windows (
+			id INTEGER PRIMARY KEY AUTOINCREMENT,
+			monitor_id INTEGER DEFAULT 0,
+			title TEXT NOT NULL,
+			description TEXT DEFAULT '',
+			type TEXT DEFAULT 'maintenance',
+			start_time DATETIME NOT NULL,
+			end_time DATETIME,
+			created_by TEXT DEFAULT '',
+			created_at DATETIME DEFAULT CURRENT_TIMESTAMP
+		)`,
 	}
 }
 
@@ -96,6 +107,8 @@ func (d *SQLiteDialect) ImportWipe(tx *sql.Tx) {
 	tx.Exec("DELETE FROM sqlite_sequence WHERE name='alerts'")
 	tx.Exec("DELETE FROM users")
 	tx.Exec("DELETE FROM sqlite_sequence WHERE name='users'")
+	tx.Exec("DELETE FROM maintenance_windows")
+	tx.Exec("DELETE FROM sqlite_sequence WHERE name='maintenance_windows'")
 }
 
 func (d *SQLiteDialect) ImportResetSequences(tx *sql.Tx) {}

internal/store/sqlstore.go

@@ -7,6 +7,7 @@ import (
 	"encoding/json"
 	"fmt"
 	"go-upkeep/internal/models"
+	"time"
 )
 
 type SQLStore struct {
@@ -356,6 +357,90 @@ func (s *SQLStore) LoadAllHistory(limit int) (map[int][]models.CheckRecord, erro
 	return result, rows.Err()
 }
 
+func (s *SQLStore) scanMaintenanceWindow(rows *sql.Rows) (models.MaintenanceWindow, error) {
+	var mw models.MaintenanceWindow
+	var endTime sql.NullTime
+	if err := rows.Scan(&mw.ID, &mw.MonitorID, &mw.Title, &mw.Description, &mw.Type, &mw.StartTime, &endTime, &mw.CreatedBy, &mw.CreatedAt); err != nil {
+		return mw, err
+	}
+	if endTime.Valid {
+		mw.EndTime = endTime.Time
+	}
+	return mw, nil
+}
+
+func (s *SQLStore) GetActiveMaintenanceWindows() ([]models.MaintenanceWindow, error) {
+	rows, err := s.db.Query(s.q("SELECT id, monitor_id, title, description, type, start_time, end_time, created_by, created_at FROM maintenance_windows WHERE start_time <= CURRENT_TIMESTAMP AND (end_time IS NULL OR end_time > CURRENT_TIMESTAMP) ORDER BY start_time DESC"))
+	if err != nil {
+		return nil, err
+	}
+	defer rows.Close()
+	var windows []models.MaintenanceWindow
+	for rows.Next() {
+		mw, err := s.scanMaintenanceWindow(rows)
+		if err != nil {
+			return windows, err
+		}
+		windows = append(windows, mw)
+	}
+	return windows, rows.Err()
+}
+
+func (s *SQLStore) GetAllMaintenanceWindows(limit int) ([]models.MaintenanceWindow, error) {
+	rows, err := s.db.Query(s.q("SELECT id, monitor_id, title, description, type, start_time, end_time, created_by, created_at FROM maintenance_windows ORDER BY created_at DESC LIMIT ?"), limit)
+	if err != nil {
+		return nil, err
+	}
+	defer rows.Close()
+	var windows []models.MaintenanceWindow
+	for rows.Next() {
+		mw, err := s.scanMaintenanceWindow(rows)
+		if err != nil {
+			return windows, err
+		}
+		windows = append(windows, mw)
+	}
+	return windows, rows.Err()
+}
+
+func (s *SQLStore) AddMaintenanceWindow(mw models.MaintenanceWindow) error {
+	if mw.StartTime.IsZero() {
+		mw.StartTime = time.Now()
+	}
+	_, err := s.db.Exec(s.q("INSERT INTO maintenance_windows (monitor_id, title, description, type, start_time, end_time, created_by) VALUES (?, ?, ?, ?, ?, ?, ?)"),
+		mw.MonitorID, mw.Title, mw.Description, mw.Type, mw.StartTime, sql.NullTime{Time: mw.EndTime, Valid: !mw.EndTime.IsZero()}, mw.CreatedBy)
+	return err
+}
+
+func (s *SQLStore) EndMaintenanceWindow(id int) error {
+	_, err := s.db.Exec(s.q("UPDATE maintenance_windows SET end_time = CURRENT_TIMESTAMP WHERE id = ?"), id)
+	return err
+}
+
+func (s *SQLStore) DeleteMaintenanceWindow(id int) error {
+	_, err := s.db.Exec(s.q("DELETE FROM maintenance_windows WHERE id = ?"), id)
+	if err != nil {
+		return err
+	}
+	s.dialect.ResetSequenceOnEmpty(s.db, "maintenance_windows")
+	return nil
+}
+
+func (s *SQLStore) IsMonitorInMaintenance(monitorID int) (bool, error) {
+	var count int
+	err := s.db.QueryRow(s.q(`SELECT COUNT(*) FROM maintenance_windows
+		WHERE type = 'maintenance'
+		AND start_time <= CURRENT_TIMESTAMP
+		AND (end_time IS NULL OR end_time > CURRENT_TIMESTAMP)
+		AND (monitor_id = 0 OR monitor_id = ?
+			OR monitor_id IN (SELECT parent_id FROM sites WHERE id = ? AND parent_id > 0))`),
+		monitorID, monitorID).Scan(&count)
+	if err != nil {
+		return false, err
+	}
+	return count > 0, nil
+}
+
 func (s *SQLStore) ExportData() (models.Backup, error) {
 	sites, err := s.GetSites()
 	if err != nil {
@@ -369,7 +454,11 @@ func (s *SQLStore) ExportData() (models.Backup, error) {
 	if err != nil {
 		return models.Backup{}, err
 	}
-	return models.Backup{Sites: sites, Alerts: alerts, Users: users}, nil
+	windows, err := s.GetAllMaintenanceWindows(1000)
+	if err != nil {
+		return models.Backup{}, err
+	}
+	return models.Backup{Sites: sites, Alerts: alerts, Users: users, MaintenanceWindows: windows}, nil
 }
 
 func (s *SQLStore) ImportData(data models.Backup) error {
@@ -403,6 +492,13 @@ func (s *SQLStore) ImportData(data models.Backup) error {
 		}
 	}
 
+	for _, mw := range data.MaintenanceWindows {
+		if _, err := tx.Exec(s.q("INSERT INTO maintenance_windows (id, monitor_id, title, description, type, start_time, end_time, created_by) VALUES (?, ?, ?, ?, ?, ?, ?, ?)"),
+			mw.ID, mw.MonitorID, mw.Title, mw.Description, mw.Type, mw.StartTime, sql.NullTime{Time: mw.EndTime, Valid: !mw.EndTime.IsZero()}, mw.CreatedBy); err != nil {
+			return err
+		}
+	}
+
 	s.dialect.ImportResetSequences(tx)
 
 	return tx.Commit()

internal/store/store.go

@@ -49,6 +49,14 @@ type Store interface {
 	SaveLog(message string) error
 	LoadLogs(limit int) ([]string, error)
 
+	// Maintenance Windows
+	GetActiveMaintenanceWindows() ([]models.MaintenanceWindow, error)
+	GetAllMaintenanceWindows(limit int) ([]models.MaintenanceWindow, error)
+	AddMaintenanceWindow(mw models.MaintenanceWindow) error
+	EndMaintenanceWindow(id int) error
+	DeleteMaintenanceWindow(id int) error
+	IsMonitorInMaintenance(monitorID int) (bool, error)
+
 	// Backup & Restore
 	ExportData() (models.Backup, error)
 	ImportData(data models.Backup) error

internal/tui/tab_maint.go

@@ -0,0 +1,230 @@
+package tui
+
+import (
+	"fmt"
+	"go-upkeep/internal/models"
+	"strconv"
+	"time"
+
+	tea "github.com/charmbracelet/bubbletea"
+	"github.com/charmbracelet/huh"
+	"github.com/charmbracelet/lipgloss"
+)
+
+var maintStyle = lipgloss.NewStyle().Foreground(lipgloss.Color("#bb9af7"))
+
+type maintFormData struct {
+	Title       string
+	Description string
+	Type        string
+	MonitorID   string
+	Duration    string
+	CustomHours string
+}
+
+func fmtMaintStatus(mw models.MaintenanceWindow) string {
+	now := time.Now()
+	if mw.StartTime.After(now) {
+		return warnStyle.Render("SCHEDULED")
+	}
+	if !mw.EndTime.IsZero() && mw.EndTime.Before(now) {
+		return subtleStyle.Render("ENDED")
+	}
+	return specialStyle.Render("ACTIVE")
+}
+
+func fmtMaintType(t string) string {
+	if t == "incident" {
+		return dangerStyle.Render("incident")
+	}
+	return maintStyle.Render("maintenance")
+}
+
+func fmtMaintMonitor(monitorID int, sites []models.Site) string {
+	if monitorID == 0 {
+		return "All"
+	}
+	for _, s := range sites {
+		if s.ID == monitorID {
+			return limitStr(s.Name, 18)
+		}
+	}
+	return fmt.Sprintf("#%d", monitorID)
+}
+
+func fmtMaintTime(t time.Time) string {
+	if t.IsZero() {
+		return subtleStyle.Render("—")
+	}
+	now := time.Now()
+	if t.Year() == now.Year() && t.YearDay() == now.YearDay() {
+		return t.Format("15:04")
+	}
+	return t.Format("15:04 Jan 02")
+}
+
+func (m Model) isMonitorInMaintenance(monitorID int) bool {
+	for _, mw := range m.maintenanceWindows {
+		if mw.Type != "maintenance" {
+			continue
+		}
+		now := time.Now()
+		if mw.StartTime.After(now) {
+			continue
+		}
+		if !mw.EndTime.IsZero() && mw.EndTime.Before(now) {
+			continue
+		}
+		if mw.MonitorID == 0 || mw.MonitorID == monitorID {
+			return true
+		}
+		for _, s := range m.sites {
+			if s.ID == monitorID && s.ParentID > 0 && mw.MonitorID == s.ParentID {
+				return true
+			}
+		}
+	}
+	return false
+}
+
+func (m Model) viewMaintTab() string {
+	if len(m.maintenanceWindows) == 0 {
+		return "\n  No maintenance windows or incidents. Press [n] to create one."
+	}
+
+	return m.renderTable(
+		[]string{"#", "TITLE", "TYPE", "MONITORS", "STATUS", "STARTED", "ENDS"},
+		len(m.maintenanceWindows),
+		func(start, end int) [][]string {
+			var rows [][]string
+			allSites := m.engine.GetAllSites()
+			for i := start; i < end; i++ {
+				mw := m.maintenanceWindows[i]
+				rows = append(rows, []string{
+					strconv.Itoa(i + 1),
+					m.zones.Mark(fmt.Sprintf("maint-%d", i), limitStr(mw.Title, 24)),
+					fmtMaintType(mw.Type),
+					fmtMaintMonitor(mw.MonitorID, allSites),
+					fmtMaintStatus(mw),
+					fmtMaintTime(mw.StartTime),
+					fmtMaintTime(mw.EndTime),
+				})
+			}
+			return rows
+		},
+		[]int{6, 0, 14, 20, 12, 16, 16},
+		nil,
+	)
+}
+
+func (m *Model) initMaintHuhForm() tea.Cmd {
+	m.maintFormData = &maintFormData{
+		Type:        "maintenance",
+		MonitorID:   "0",
+		Duration:    "1h",
+		CustomHours: "12",
+	}
+
+	monitorOpts := []huh.Option[string]{huh.NewOption("All Monitors", "0")}
+	allSites := m.engine.GetAllSites()
+	for _, s := range allSites {
+		label := s.Name
+		if s.Type == "group" {
+			label = s.Name + " (group)"
+		}
+		monitorOpts = append(monitorOpts, huh.NewOption(label, strconv.Itoa(s.ID)))
+	}
+
+	m.huhForm = huh.NewForm(
+		huh.NewGroup(
+			huh.NewInput().Title("Title").
+				Placeholder("DB Migration").
+				Value(&m.maintFormData.Title).
+				Validate(func(s string) error {
+					if s == "" {
+						return fmt.Errorf("title is required")
+					}
+					return nil
+				}),
+			huh.NewSelect[string]().Title("Type").
+				Options(
+					huh.NewOption("Maintenance (suppress alerts)", "maintenance"),
+					huh.NewOption("Incident (informational)", "incident"),
+				).Value(&m.maintFormData.Type),
+			huh.NewSelect[string]().Title("Affected Monitors").
+				Options(monitorOpts...).
+				Value(&m.maintFormData.MonitorID),
+			huh.NewInput().Title("Description").
+				Placeholder("Optional notes").
+				Value(&m.maintFormData.Description),
+		).Title("Maintenance Window"),
+		huh.NewGroup(
+			huh.NewSelect[string]().Title("Duration").
+				Options(
+					huh.NewOption("1 hour", "1h"),
+					huh.NewOption("2 hours", "2h"),
+					huh.NewOption("4 hours", "4h"),
+					huh.NewOption("8 hours", "8h"),
+					huh.NewOption("Indefinite (end manually)", "indefinite"),
+					huh.NewOption("Custom", "custom"),
+				).Value(&m.maintFormData.Duration),
+			huh.NewInput().Title("Custom Duration (hours)").
+				Placeholder("12").
+				Value(&m.maintFormData.CustomHours).
+				Validate(func(s string) error {
+					if m.maintFormData.Duration != "custom" {
+						return nil
+					}
+					v, err := strconv.Atoi(s)
+					if err != nil {
+						return fmt.Errorf("must be a number")
+					}
+					if v < 1 {
+						return fmt.Errorf("must be at least 1 hour")
+					}
+					return nil
+				}),
+		).Title("Duration").WithHideFunc(func() bool {
+			return m.maintFormData.Type == "incident"
+		}),
+	).WithTheme(huh.ThemeDracula())
+
+	return m.huhForm.Init()
+}
+
+func (m *Model) submitMaintForm() {
+	d := m.maintFormData
+	monitorID, _ := strconv.Atoi(d.MonitorID)
+
+	mw := models.MaintenanceWindow{
+		MonitorID:   monitorID,
+		Title:       d.Title,
+		Description: d.Description,
+		Type:        d.Type,
+		StartTime:   time.Now(),
+	}
+
+	if d.Type == "maintenance" {
+		switch d.Duration {
+		case "1h":
+			mw.EndTime = mw.StartTime.Add(1 * time.Hour)
+		case "2h":
+			mw.EndTime = mw.StartTime.Add(2 * time.Hour)
+		case "4h":
+			mw.EndTime = mw.StartTime.Add(4 * time.Hour)
+		case "8h":
+			mw.EndTime = mw.StartTime.Add(8 * time.Hour)
+		case "custom":
+			hours, _ := strconv.Atoi(d.CustomHours)
+			if hours < 1 {
+				hours = 1
+			}
+			mw.EndTime = mw.StartTime.Add(time.Duration(hours) * time.Hour)
+		}
+	}
+
+	if err := m.store.AddMaintenanceWindow(mw); err != nil {
+		m.engine.AddLog("Add maintenance window failed: " + err.Error())
+	}
+	m.state = stateDashboard
+}

internal/tui/tab_sites.go

@@ -207,10 +207,13 @@ func fmtRetries(site models.Site) string {
 	return s
 }
 
-func fmtStatus(status string, paused bool) string {
+func fmtStatus(status string, paused bool, inMaint bool) string {
 	if paused {
 		return warnStyle.Render("PAUSED")
 	}
+	if inMaint {
+		return maintStyle.Render("MAINT")
+	}
 	switch {
 	case status == "DOWN" || status == "SSL EXP":
 		return dangerStyle.Render(status)
@@ -280,7 +283,7 @@ func (m Model) viewSitesTab() string {
 						strconv.Itoa(i + 1),
 						m.zones.Mark(fmt.Sprintf("site-%d", i), icon+" "+limitStr(site.Name, nameW-2)),
 						"group",
-						fmtStatus(site.Status, site.Paused),
+						fmtStatus(site.Status, site.Paused, m.isMonitorInMaintenance(site.ID)),
 						subtleStyle.Render("—"),
 						subtleStyle.Render("—"),
 						subtleStyle.Render(strings.Repeat("·", sparkWidth)),
@@ -313,7 +316,7 @@ func (m Model) viewSitesTab() string {
 					strconv.Itoa(i + 1),
 					m.zones.Mark(fmt.Sprintf("site-%d", i), name),
 					typeIcon(site.Type, false) + " " + site.Type,
-					fmtStatus(site.Status, site.Paused),
+					fmtStatus(site.Status, site.Paused, m.isMonitorInMaintenance(site.ID)),
 					fmtLatency(site.Latency),
 					fmtUptime(hist.Statuses),
 					spark,
@@ -623,7 +626,15 @@ func (m Model) viewDetailPanel() string {
 		b.WriteString(fmt.Sprintf("  %-16s %s\n", subtleStyle.Render(label), value))
 	}
 
-	row("Status", fmtStatus(site.Status, site.Paused))
+	row("Status", fmtStatus(site.Status, site.Paused, m.isMonitorInMaintenance(site.ID)))
+	if m.isMonitorInMaintenance(site.ID) {
+		for _, mw := range m.maintenanceWindows {
+			if mw.Type == "maintenance" && (mw.MonitorID == 0 || mw.MonitorID == site.ID || mw.MonitorID == site.ParentID) {
+				row("Maintenance", maintStyle.Render(mw.Title))
+				break
+			}
+		}
+	}
 	row("Type", site.Type)
 	if site.URL != "" {
 		row("URL", site.URL)

internal/tui/tui.go

@@ -42,6 +42,7 @@ const (
 	stateFormAlert
 	stateFormUser
 	stateConfirmDelete
+	stateFormMaint
 )
 
 type Model struct {
@@ -59,6 +60,7 @@ type Model struct {
 	siteFormData  *siteFormData
 	alertFormData *alertFormData
 	userFormData  *userFormData
+	maintFormData *maintFormData
 
 	logViewport viewport.Model
 	isAdmin     bool
@@ -78,10 +80,11 @@ type Model struct {
 	pulseVel    float64
 	tickCount   int
 
-	sites  []models.Site
-	alerts []models.AlertConfig
-	users  []models.User
-	nodes  []models.ProbeNode
+	sites              []models.Site
+	alerts             []models.AlertConfig
+	users              []models.User
+	nodes              []models.ProbeNode
+	maintenanceWindows []models.MaintenanceWindow
 
 	filterMode bool
 	filterText string
@@ -128,7 +131,12 @@ func (m Model) Update(msg tea.Msg) (tea.Model, tea.Cmd) {
 						m.engine.AddLog("Delete alert failed: " + err.Error())
 					}
 					m.adjustCursor(len(m.alerts) - 1)
-				case 3:
+				case 4:
+					if err := m.store.DeleteMaintenanceWindow(m.deleteID); err != nil {
+						m.engine.AddLog("Delete maintenance window failed: " + err.Error())
+					}
+					m.adjustCursor(len(m.maintenanceWindows) - 1)
+				case 5:
 					if err := m.store.DeleteUser(m.deleteID); err != nil {
 						m.engine.AddLog("Delete user failed: " + err.Error())
 					}
@@ -136,12 +144,12 @@ func (m Model) Update(msg tea.Msg) (tea.Model, tea.Cmd) {
 				}
 				m.refreshData()
 				m.state = stateDashboard
-				if m.deleteTab == 4 {
+				if m.deleteTab == 5 {
 					m.state = stateUsers
 				}
 			case "n", "N", "esc":
 				m.state = stateDashboard
-				if m.deleteTab == 4 {
+				if m.deleteTab == 5 {
 					m.state = stateUsers
 				}
 			case "ctrl+c":
@@ -152,7 +160,11 @@ func (m Model) Update(msg tea.Msg) (tea.Model, tea.Cmd) {
 	}
 
 	// Form state: forward ALL messages to huh (keys, timers, resize, etc.)
-	if m.state == stateFormSite || m.state == stateFormAlert || m.state == stateFormUser {
+	if m.state == stateFormSite || m.state == stateFormAlert || m.state == stateFormUser || m.state == stateFormMaint {
+		if wsm, ok := msg.(tea.WindowSizeMsg); ok {
+			m.termWidth = wsm.Width
+			m.termHeight = wsm.Height
+		}
 		if keyMsg, ok := msg.(tea.KeyMsg); ok {
 			if keyMsg.String() == "ctrl+c" {
 				return m, tea.Quit
@@ -160,7 +172,7 @@ func (m Model) Update(msg tea.Msg) (tea.Model, tea.Cmd) {
 			if keyMsg.String() == "esc" {
 				m.huhForm = nil
 				m.state = stateDashboard
-				if m.currentTab == 4 {
+				if m.currentTab == 5 {
 					m.state = stateUsers
 				}
 				return m, nil
@@ -226,6 +238,8 @@ func (m Model) Update(msg tea.Msg) (tea.Model, tea.Cmd) {
 				} else if m.currentTab == 3 {
 					listLen = len(m.nodes)
 				} else if m.currentTab == 4 {
+					listLen = len(m.maintenanceWindows)
+				} else if m.currentTab == 5 {
 					listLen = len(m.users)
 				}
 				if msg.Button == tea.MouseButtonWheelUp {
@@ -331,6 +345,9 @@ func (m Model) Update(msg tea.Msg) (tea.Model, tea.Cmd) {
 						max = len(m.nodes) - 1
 					}
 					if m.currentTab == 4 {
+						max = len(m.maintenanceWindows) - 1
+					}
+					if m.currentTab == 5 {
 						max = len(m.users) - 1
 					}
 					if m.cursor < max {
@@ -349,7 +366,10 @@ func (m Model) Update(msg tea.Msg) (tea.Model, tea.Cmd) {
 				} else if m.currentTab == 1 {
 					m.state = stateFormAlert
 					return m, m.initAlertHuhForm()
-				} else if m.currentTab == 4 && m.isAdmin {
+				} else if m.currentTab == 4 {
+					m.state = stateFormMaint
+					return m, m.initMaintHuhForm()
+				} else if m.currentTab == 5 && m.isAdmin {
 					m.state = stateFormUser
 					return m, m.initUserHuhForm()
 				}
@@ -363,7 +383,7 @@ func (m Model) Update(msg tea.Msg) (tea.Model, tea.Cmd) {
 					m.editID = m.alerts[m.cursor].ID
 					m.state = stateFormAlert
 					return m, m.initAlertHuhForm()
-				} else if m.currentTab == 4 && m.isAdmin && len(m.users) > 0 {
+				} else if m.currentTab == 5 && m.isAdmin && len(m.users) > 0 {
 					m.editID = m.users[m.cursor].ID
 					m.state = stateFormUser
 					return m, m.initUserHuhForm()
@@ -386,6 +406,18 @@ func (m Model) Update(msg tea.Msg) (tea.Model, tea.Cmd) {
 				if m.currentTab == 0 && len(m.sites) > 0 {
 					m.state = stateDetail
 				}
+			case "x":
+				if m.currentTab == 4 && len(m.maintenanceWindows) > 0 {
+					mw := m.maintenanceWindows[m.cursor]
+					now := time.Now()
+					isActive := !mw.StartTime.After(now) && (mw.EndTime.IsZero() || mw.EndTime.After(now))
+					if isActive {
+						if err := m.store.EndMaintenanceWindow(mw.ID); err != nil {
+							m.engine.AddLog("End maintenance failed: " + err.Error())
+						}
+						m.refreshData()
+					}
+				}
 			case "d", "backspace":
 				if m.currentTab == 0 && len(m.sites) > 0 {
 					m.deleteID = m.sites[m.cursor].ID
@@ -397,10 +429,15 @@ func (m Model) Update(msg tea.Msg) (tea.Model, tea.Cmd) {
 					m.deleteName = m.alerts[m.cursor].Name
 					m.deleteTab = 1
 					m.state = stateConfirmDelete
-				} else if m.currentTab == 4 && m.isAdmin && len(m.users) > 0 {
+				} else if m.currentTab == 4 && len(m.maintenanceWindows) > 0 {
+					m.deleteID = m.maintenanceWindows[m.cursor].ID
+					m.deleteName = m.maintenanceWindows[m.cursor].Title
+					m.deleteTab = 4
+					m.state = stateConfirmDelete
+				} else if m.currentTab == 5 && m.isAdmin && len(m.users) > 0 {
 					m.deleteID = m.users[m.cursor].ID
 					m.deleteName = m.users[m.cursor].Username
-					m.deleteTab = 4
+					m.deleteTab = 5
 					m.state = stateConfirmDelete
 				}
 			}
@@ -410,9 +447,9 @@ func (m Model) Update(msg tea.Msg) (tea.Model, tea.Cmd) {
 }
 
 func (m *Model) handleClick(msg tea.MouseMsg) (tea.Model, tea.Cmd) {
-	tabCount := 4
+	tabCount := 5
 	if m.isAdmin {
-		tabCount = 5
+		tabCount = 6
 	}
 	for i := 0; i < tabCount; i++ {
 		if m.zones.Get(fmt.Sprintf("tab-%d", i)).InBounds(msg) {
@@ -448,6 +485,19 @@ func (m *Model) handleClick(msg tea.MouseMsg) (tea.Model, tea.Cmd) {
 	}
 
 	if m.currentTab == 4 {
+		end := m.tableOffset + m.maxTableRows
+		if end > len(m.maintenanceWindows) {
+			end = len(m.maintenanceWindows)
+		}
+		for i := m.tableOffset; i < end; i++ {
+			if m.zones.Get(fmt.Sprintf("maint-%d", i)).InBounds(msg) {
+				m.cursor = i
+				return m, nil
+			}
+		}
+	}
+
+	if m.currentTab == 5 {
 		end := m.tableOffset + m.maxTableRows
 		if end > len(m.users) {
 			end = len(m.users)
@@ -464,9 +514,9 @@ func (m *Model) handleClick(msg tea.MouseMsg) (tea.Model, tea.Cmd) {
 }
 
 func (m *Model) switchTab(idx int) {
-	maxTabs := 3
+	maxTabs := 4
 	if m.isAdmin {
-		maxTabs = 4
+		maxTabs = 5
 	}
 	if idx > maxTabs {
 		idx = 0
@@ -477,7 +527,7 @@ func (m *Model) switchTab(idx int) {
 	switch idx {
 	case 2:
 		m.state = stateLogs
-	case 4:
+	case 5:
 		m.state = stateUsers
 	default:
 		m.state = stateDashboard
@@ -550,6 +600,9 @@ func (m *Model) refreshData() {
 	if nodes, err := m.store.GetAllNodes(); err == nil {
 		m.nodes = nodes
 	}
+	if windows, err := m.store.GetAllMaintenanceWindows(100); err == nil {
+		m.maintenanceWindows = windows
+	}
 	m.logViewport.SetContent(strings.Join(m.engine.GetLogs(), "\n"))
 
 	listLen := len(m.sites)
@@ -558,6 +611,8 @@ func (m *Model) refreshData() {
 	} else if m.currentTab == 3 {
 		listLen = len(m.nodes)
 	} else if m.currentTab == 4 {
+		listLen = len(m.maintenanceWindows)
+	} else if m.currentTab == 5 {
 		listLen = len(m.users)
 	}
 	if listLen > 0 && m.cursor >= listLen {
@@ -582,6 +637,10 @@ func (m *Model) submitForm() {
 		if m.userFormData != nil {
 			m.submitUserForm()
 		}
+	case stateFormMaint:
+		if m.maintFormData != nil {
+			m.submitMaintForm()
+		}
 	}
 }
 
@@ -593,7 +652,7 @@ func (m Model) pulseIndicator() string {
 	}
 	hasDown := false
 	for _, s := range m.sites {
-		if !s.Paused && (s.Status == "DOWN" || s.Status == "SSL EXP") {
+		if !s.Paused && !m.isMonitorInMaintenance(s.ID) && (s.Status == "DOWN" || s.Status == "SSL EXP") {
 			hasDown = true
 			break
 		}
@@ -614,6 +673,8 @@ func (m Model) View() string {
 		if m.deleteTab == 1 {
 			kind = "alert"
 		} else if m.deleteTab == 4 {
+			kind = "maintenance window"
+		} else if m.deleteTab == 5 {
 			kind = "user"
 		}
 		msg := dangerStyle.Render(fmt.Sprintf("Delete %s \"%s\"?", kind, m.deleteName))
@@ -624,7 +685,7 @@ func (m Model) View() string {
 			Padding(1, 3).
 			Render(msg + "\n\n" + hint)
 		return lipgloss.NewStyle().Padding(2, 4).Render(box)
-	case stateFormSite, stateFormAlert, stateFormUser:
+	case stateFormSite, stateFormAlert, stateFormUser, stateFormMaint:
 		if m.huhForm != nil {
 			title := ""
 			switch m.state {
@@ -643,7 +704,14 @@ func (m Model) View() string {
 				if m.editID > 0 {
 					title = fmt.Sprintf("Edit User #%d", m.editID)
 				}
+			case stateFormMaint:
+				title = "New Maintenance Window"
 			}
+			formHeight := m.termHeight - 7
+			if formHeight < 5 {
+				formHeight = 5
+			}
+			m.huhForm.WithHeight(formHeight)
 			header := titleStyle.Render(title)
 			footer := subtleStyle.Render("\n[Esc] Cancel")
 			return lipgloss.NewStyle().Padding(1, 2).Render(header + "\n\n" + m.huhForm.View() + "\n" + footer)
@@ -659,7 +727,7 @@ func (m Model) View() string {
 func (m Model) viewDashboard() string {
 	downCount := 0
 	for _, s := range m.sites {
-		if !s.Paused && (s.Status == "DOWN" || s.Status == "SSL EXP") {
+		if !s.Paused && !m.isMonitorInMaintenance(s.ID) && (s.Status == "DOWN" || s.Status == "SSL EXP") {
 			downCount++
 		}
 	}
@@ -687,7 +755,21 @@ func (m Model) viewDashboard() string {
 		nodesLabel = "Nodes"
 	}
 
-	tabs := []string{sitesLabel, "Alerts", "Logs", nodesLabel}
+	activeMaint := 0
+	for _, mw := range m.maintenanceWindows {
+		now := time.Now()
+		if !mw.StartTime.After(now) && (mw.EndTime.IsZero() || mw.EndTime.After(now)) {
+			activeMaint++
+		}
+	}
+	var maintLabel string
+	if activeMaint > 0 {
+		maintLabel = fmt.Sprintf("Maint (%d)", activeMaint)
+	} else {
+		maintLabel = "Maint"
+	}
+
+	tabs := []string{sitesLabel, "Alerts", "Logs", nodesLabel, maintLabel}
 	if m.isAdmin {
 		tabs = append(tabs, "Users")
 	}
@@ -717,6 +799,8 @@ func (m Model) viewDashboard() string {
 	case 3:
 		content = m.viewNodesTab()
 	case 4:
+		content = m.viewMaintTab()
+	case 5:
 		if m.isAdmin {
 			content = m.viewUsersTab()
 		}
@@ -751,6 +835,8 @@ func (m Model) viewDashboard() string {
 		case 0:
 			keys = "[/]Filter [n]New [e]Edit [i]Info [d]Del [p]Pause [Tab]Switch [q]Quit"
 		case 4:
+			keys = "[n]New [x]End [d]Del [Tab]Switch [q]Quit"
+		case 5:
 			keys = "[n]Add [d]Revoke [Tab]Switch [q]Quit"
 		default:
 			keys = "[Tab]Switch [q]Quit"