chore: add TUI screenshots

Six tab screenshots: monitors, alerts, logs, nodes, detail, theme. Generated with VHS; tooling lives in lerko/uptop-vhs.
feat(tui): add detail panel, type icons, and store improvements
2026-05-29 12:35:15 -04:00 · 2026-05-29 12:35:08 -04:00 · 2026-05-28 20:40:29 +00:00 · 2026-05-28 01:33:00 +00:00 · 2026-05-27 21:23:06 -04:00 · 2026-05-28 00:35:24 +00:00
31 changed files with 1065 additions and 239 deletions
@@ -27,7 +27,7 @@ go.work
 # End of https://www.toptal.com/developers/gitignore/api/go

 /uptop
-uptop.db
+uptop.db*

 .ssh

@@ -385,6 +385,7 @@ func runServe(args []string) {

 	eng.InitHistory()
 	eng.InitLogs()
+	eng.InitAlertHealth()
 	eng.Start(ctx)

 	tlsCert := os.Getenv("UPTOP_TLS_CERT")
@@ -53,21 +53,27 @@ func (m *mockStore) GetNode(string) (models.ProbeNode, error) { return models.Pr
 func (m *mockStore) GetAllNodes() ([]models.ProbeNode, error) { return nil, nil }
 func (m *mockStore) UpdateNodeLastSeen(string) error          { return nil }
 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) LoadAlertHealth() (map[int]models.AlertHealthRecord, error) {
+	return nil, nil
+}
+func (m *mockStore) SaveAlertHealth(models.AlertHealthRecord) 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 (m *mockStore) GetPreference(string) (string, error)                { return "", nil }
-func (m *mockStore) SetPreference(string, string) error                  { return nil }
-func (m *mockStore) Close() error                                        { return 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 (m *mockStore) GetPreference(string) (string, error)                   { return "", nil }
+func (m *mockStore) SetPreference(string, string) error                     { return nil }
+func (m *mockStore) SaveStateChange(int, string, string, string) error      { return nil }
+func (m *mockStore) GetStateChanges(int, int) ([]models.StateChange, error) { return nil, nil }
+func (m *mockStore) Close() error                                           { return nil }

 // --- Cluster Start Tests ---

@@ -127,9 +127,10 @@ func probeFetchAssignments(ctx context.Context, client *http.Client, cfg ProbeCo
 }

 type probeResultItem struct {
-	SiteID    int   `json:"site_id"`
-	LatencyNs int64 `json:"latency_ns"`
-	IsUp      bool  `json:"is_up"`
+	SiteID      int    `json:"site_id"`
+	LatencyNs   int64  `json:"latency_ns"`
+	IsUp        bool   `json:"is_up"`
+	ErrorReason string `json:"error_reason,omitempty"`
 }

 func probeExecuteChecks(ctx context.Context, sites []models.Site, strict, insecure *http.Client, allowPrivate bool) []probeResultItem {
@@ -154,9 +155,10 @@ loop:
 			cr := monitor.RunCheck(s, strict, insecure, false, allowPrivate)
 			mu.Lock()
 			results = append(results, probeResultItem{
-				SiteID:    s.ID,
-				LatencyNs: cr.LatencyNs,
-				IsUp:      cr.Status == "UP",
+				SiteID:      s.ID,
+				LatencyNs:   cr.LatencyNs,
+				IsUp:        cr.Status == "UP",
+				ErrorReason: cr.ErrorReason,
 			})
 			mu.Unlock()
 		}(site)
@@ -2,13 +2,14 @@ package metrics

 import (
 	"context"
-	"gitea.lerkolabs.com/lerko/uptop/internal/models"
-	"gitea.lerkolabs.com/lerko/uptop/internal/monitor"
 	"net/http"
 	"net/http/httptest"
 	"strings"
 	"testing"
 	"time"
+
+	"gitea.lerkolabs.com/lerko/uptop/internal/models"
+	"gitea.lerkolabs.com/lerko/uptop/internal/monitor"
 )

 type mockStore struct {
@@ -50,21 +51,27 @@ func (m *mockStore) GetNode(string) (models.ProbeNode, error)         { return m
 func (m *mockStore) GetAllNodes() ([]models.ProbeNode, error)         { return nil, nil }
 func (m *mockStore) UpdateNodeLastSeen(string) error                  { return nil }
 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) LoadAlertHealth() (map[int]models.AlertHealthRecord, error) {
+	return nil, nil
+}
+func (m *mockStore) SaveAlertHealth(models.AlertHealthRecord) 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 (m *mockStore) GetPreference(string) (string, error)                { return "", nil }
-func (m *mockStore) SetPreference(string, string) error                  { return nil }
-func (m *mockStore) Close() error                                        { return 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 (m *mockStore) GetPreference(string) (string, error)                   { return "", nil }
+func (m *mockStore) SetPreference(string, string) error                     { return nil }
+func (m *mockStore) SaveStateChange(int, string, string, string) error      { return nil }
+func (m *mockStore) GetStateChanges(int, int) ([]models.StateChange, error) { return nil, nil }
+func (m *mockStore) Close() error                                           { return nil }

 func TestMetricsHandler(t *testing.T) {
 	ms := &mockStore{
@@ -27,14 +27,26 @@ type Site struct {
 	Paused         bool
 	Regions        string

-	FailureCount   int
-	Status         string
-	StatusCode     int
-	Latency        time.Duration
-	CertExpiry     time.Time
-	HasSSL         bool
-	LastCheck      time.Time
-	SentSSLWarning bool
+	FailureCount    int
+	Status          string
+	StatusCode      int
+	Latency         time.Duration
+	CertExpiry      time.Time
+	HasSSL          bool
+	LastCheck       time.Time
+	SentSSLWarning  bool
+	LastError       string
+	StatusChangedAt time.Time
+	LastSuccessAt   time.Time
+}
+
+type StateChange struct {
+	ID          int
+	SiteID      int
+	FromStatus  string
+	ToStatus    string
+	ErrorReason string
+	ChangedAt   time.Time
 }

 type AlertConfig struct {
@@ -67,6 +79,17 @@ type ProbeNode struct {
 	Version  string
 }

+// AlertHealthRecord is the persisted send health of an alert channel. It lets the
+// "last sent" / health indicators survive restarts instead of resetting to "never".
+type AlertHealthRecord struct {
+	AlertID    int
+	LastSendAt time.Time
+	LastSendOK bool
+	LastError  string
+	SendCount  int
+	FailCount  int
+}
+
 type MaintenanceWindow struct {
 	ID          int
 	MonitorID   int
@@ -11,10 +11,11 @@ const (
 )

 type NodeResult struct {
-	NodeID    string
-	IsUp      bool
-	LatencyNs int64
-	CheckedAt time.Time
+	NodeID      string
+	IsUp        bool
+	LatencyNs   int64
+	CheckedAt   time.Time
+	ErrorReason string
 }

 func AggregateStatus(results []NodeResult, strategy AggregationStrategy) (isUp bool, avgLatencyNs int64) {
@@ -2,6 +2,7 @@ package monitor

 import (
 	"context"
+	"fmt"
 	"net"
 	"net/http"
 	"strconv"
@@ -15,12 +16,13 @@ import (
 )

 type CheckResult struct {
-	SiteID     int
-	Status     string // "UP", "DOWN", "SSL EXP"
-	StatusCode int
-	LatencyNs  int64
-	HasSSL     bool
-	CertExpiry time.Time
+	SiteID      int
+	Status      string // "UP", "DOWN", "SSL EXP"
+	StatusCode  int
+	LatencyNs   int64
+	HasSSL      bool
+	CertExpiry  time.Time
+	ErrorReason string
 }

 func RunCheck(site models.Site, strict, insecure *http.Client, globalInsecure bool, allowPrivate ...bool) CheckResult {
@@ -35,7 +37,7 @@ func RunCheck(site models.Site, strict, insecure *http.Client, globalInsecure bo
 			if ips, err := net.LookupIP(host); err == nil {
 				for _, ip := range ips {
 					if isPrivateIP(ip) {
-						return CheckResult{SiteID: site.ID, Status: "DOWN"}
+						return CheckResult{SiteID: site.ID, Status: "DOWN", ErrorReason: "target resolves to private IP"}
 					}
 				}
 			}
@@ -52,7 +54,7 @@ func RunCheck(site models.Site, strict, insecure *http.Client, globalInsecure bo
 	case "dns":
 		return runDNSCheck(site)
 	default:
-		return CheckResult{SiteID: site.ID, Status: "DOWN"}
+		return CheckResult{SiteID: site.ID, Status: "DOWN", ErrorReason: "unsupported monitor type: " + site.Type}
 	}
 }

@@ -68,7 +70,7 @@ func runHTTPCheck(site models.Site, strict, insecure *http.Client, globalInsecur

 	req, err := http.NewRequestWithContext(ctx, method, site.URL, nil)
 	if err != nil {
-		return CheckResult{SiteID: site.ID, Status: "DOWN"}
+		return CheckResult{SiteID: site.ID, Status: "DOWN", ErrorReason: "invalid request: " + err.Error()}
 	}

 	client := strict
@@ -88,6 +90,7 @@ func runHTTPCheck(site models.Site, strict, insecure *http.Client, globalInsecur

 	if err != nil {
 		result.Status = "DOWN"
+		result.ErrorReason = truncateError(err.Error(), 256)
 		return result
 	}
 	defer resp.Body.Close()
@@ -95,6 +98,11 @@ func runHTTPCheck(site models.Site, strict, insecure *http.Client, globalInsecur
 	result.StatusCode = resp.StatusCode
 	if !isCodeAccepted(resp.StatusCode, site.AcceptedCodes) {
 		result.Status = "DOWN"
+		expected := site.AcceptedCodes
+		if expected == "" {
+			expected = "200-299"
+		}
+		result.ErrorReason = fmt.Sprintf("HTTP %d (expected %s)", resp.StatusCode, expected)
 	}

 	if site.CheckSSL && resp.TLS != nil && len(resp.TLS.PeerCertificates) > 0 {
@@ -103,6 +111,7 @@ func runHTTPCheck(site models.Site, strict, insecure *http.Client, globalInsecur
 		result.CertExpiry = cert.NotAfter
 		if time.Now().After(cert.NotAfter) {
 			result.Status = "SSL EXP"
+			result.ErrorReason = "SSL certificate expired"
 		}
 	}

@@ -117,7 +126,7 @@ func runPingCheck(site models.Site) CheckResult {

 	pinger, err := probing.NewPinger(host)
 	if err != nil {
-		return CheckResult{SiteID: site.ID, Status: "DOWN"}
+		return CheckResult{SiteID: site.ID, Status: "DOWN", ErrorReason: "ping setup: " + err.Error()}
 	}
 	pinger.Count = 1
 	pinger.Timeout = siteTimeout(site)
@@ -127,8 +136,11 @@ func runPingCheck(site models.Site) CheckResult {
 	err = pinger.Run()
 	latency := time.Since(start)

-	if err != nil || pinger.Statistics().PacketsRecv == 0 {
-		return CheckResult{SiteID: site.ID, Status: "DOWN", LatencyNs: latency.Nanoseconds()}
+	if err != nil {
+		return CheckResult{SiteID: site.ID, Status: "DOWN", LatencyNs: latency.Nanoseconds(), ErrorReason: "ping failed: " + err.Error()}
+	}
+	if pinger.Statistics().PacketsRecv == 0 {
+		return CheckResult{SiteID: site.ID, Status: "DOWN", LatencyNs: latency.Nanoseconds(), ErrorReason: "no ICMP response"}
 	}

 	stats := pinger.Statistics()
@@ -148,7 +160,7 @@ func runPortCheck(site models.Site) CheckResult {
 	latency := time.Since(start)

 	if err != nil {
-		return CheckResult{SiteID: site.ID, Status: "DOWN", LatencyNs: latency.Nanoseconds()}
+		return CheckResult{SiteID: site.ID, Status: "DOWN", LatencyNs: latency.Nanoseconds(), ErrorReason: truncateError(err.Error(), 256)}
 	}
 	_ = conn.Close()
 	return CheckResult{SiteID: site.ID, Status: "UP", LatencyNs: latency.Nanoseconds()}
@@ -199,10 +211,10 @@ func runDNSCheck(site models.Site) CheckResult {
 	latency := time.Since(start)

 	if err != nil {
-		return CheckResult{SiteID: site.ID, Status: "DOWN", LatencyNs: latency.Nanoseconds()}
+		return CheckResult{SiteID: site.ID, Status: "DOWN", LatencyNs: latency.Nanoseconds(), ErrorReason: "DNS query failed: " + err.Error()}
 	}
 	if r.Rcode != dns.RcodeSuccess {
-		return CheckResult{SiteID: site.ID, Status: "DOWN", StatusCode: r.Rcode, LatencyNs: latency.Nanoseconds()}
+		return CheckResult{SiteID: site.ID, Status: "DOWN", StatusCode: r.Rcode, LatencyNs: latency.Nanoseconds(), ErrorReason: "DNS RCODE: " + dns.RcodeToString[r.Rcode]}
 	}
 	return CheckResult{SiteID: site.ID, Status: "UP", LatencyNs: latency.Nanoseconds()}
 }
@@ -235,3 +247,10 @@ func isCodeAccepted(code int, accepted string) bool {
 	}
 	return false
 }
+
+func truncateError(s string, max int) string {
+	if len(s) <= max {
+		return s
+	}
+	return s[:max-3] + "..."
+}
@@ -19,10 +19,18 @@ import (
 const (
 	maxLogEntries    = 100
 	pollInterval     = 5 * time.Second
-	pushGracePeriod  = 5 * time.Second
 	minCheckInterval = 5
+	minPushGrace     = 60 * time.Second
 )

+type AlertHealth struct {
+	LastSendAt time.Time
+	LastSendOK bool
+	LastError  string
+	SendCount  int
+	FailCount  int
+}
+
 type Engine struct {
 	mu        sync.RWMutex
 	liveState map[int]models.Site
@@ -42,6 +50,9 @@ type Engine struct {
 	probeResults   map[int]map[string]NodeResult
 	aggStrategy    AggregationStrategy

+	alertHealthMu sync.RWMutex
+	alertHealth   map[int]AlertHealth
+
 	db                  store.Store
 	insecureSkipVerify  bool
 	allowPrivateTargets bool
@@ -64,6 +75,7 @@ func newEngine(s store.Store, allowPrivateTargets bool) *Engine {
 		histories:           make(map[int]*SiteHistory),
 		tokenIndex:          make(map[string]int),
 		probeResults:        make(map[int]map[string]NodeResult),
+		alertHealth:         make(map[int]AlertHealth),
 		aggStrategy:         AggAnyDown,
 		isActive:            true,
 		allowPrivateTargets: allowPrivateTargets,
@@ -96,6 +108,19 @@ func sanitizeLog(s string) string {
 	return s
 }

+func fmtDurationShort(d time.Duration) string {
+	if d < time.Minute {
+		return fmt.Sprintf("%ds", int(d.Seconds()))
+	}
+	if d < time.Hour {
+		return fmt.Sprintf("%dm", int(d.Minutes()))
+	}
+	if d < 24*time.Hour {
+		return fmt.Sprintf("%dh %dm", int(d.Hours()), int(d.Minutes())%60)
+	}
+	return fmt.Sprintf("%dd %dh", int(d.Hours())/24, int(d.Hours())%24)
+}
+
 func (e *Engine) AddLog(msg string) {
 	e.logMu.Lock()
 	defer e.logMu.Unlock()
@@ -121,6 +146,26 @@ func (e *Engine) InitLogs() {
 	e.logStore = logs
 }

+// InitAlertHealth restores persisted alert send health so the dashboard shows real
+// "last sent" / health state on startup instead of resetting every channel to "never".
+func (e *Engine) InitAlertHealth() {
+	records, err := e.db.LoadAlertHealth()
+	if err != nil {
+		return
+	}
+	e.alertHealthMu.Lock()
+	defer e.alertHealthMu.Unlock()
+	for id, r := range records {
+		e.alertHealth[id] = AlertHealth{
+			LastSendAt: r.LastSendAt,
+			LastSendOK: r.LastSendOK,
+			LastError:  r.LastError,
+			SendCount:  r.SendCount,
+			FailCount:  r.FailCount,
+		}
+	}
+}
+
 func (e *Engine) GetLogs() []string {
 	e.logMu.RLock()
 	defer e.logMu.RUnlock()
@@ -186,17 +231,38 @@ func (e *Engine) RecordHeartbeat(token string) bool {
 		return false
 	}

+	prevStatus := site.Status
 	site.LastCheck = time.Now()
-	wasDown := site.Status == "DOWN"
 	site.Status = "UP"
 	site.FailureCount = 0
 	site.Latency = 0
+	site.LastError = ""
+	site.LastSuccessAt = time.Now()
+
+	if prevStatus != "UP" {
+		site.StatusChangedAt = time.Now()
+	}
+
 	e.liveState[targetID] = site

-	if wasDown {
-		e.AddLog(fmt.Sprintf("Push Monitor '%s' recovered", site.Name))
-		e.triggerAlert(site.AlertID, "✅ RECOVERY", fmt.Sprintf("Push Monitor '%s' is receiving heartbeats.", site.Name))
+	switch prevStatus {
+	case "PENDING":
+		e.AddLog(fmt.Sprintf("Push Monitor '%s' received first heartbeat", site.Name))
+	case "LATE":
+		e.AddLog(fmt.Sprintf("Push Monitor '%s' heartbeat arrived (was late)", site.Name))
+	case "DOWN":
+		downDur := ""
+		if !site.StatusChangedAt.IsZero() {
+			downDur = fmt.Sprintf(" (was down %s)", fmtDurationShort(time.Since(site.StatusChangedAt)))
+		}
+		e.AddLog(fmt.Sprintf("Push Monitor '%s' recovered%s", site.Name, downDur))
+		go e.triggerAlert(site.AlertID, "✅ RECOVERY", fmt.Sprintf("Push Monitor '%s' is receiving heartbeats.%s", site.Name, downDur))
 	}
+
+	if prevStatus != "UP" && prevStatus != "PENDING" {
+		go func() { _ = e.db.SaveStateChange(targetID, prevStatus, "UP", "") }()
+	}
+
 	return true
 }

@@ -241,9 +307,6 @@ func (e *Engine) Start(ctx context.Context) {
 				if !exists {
 					e.mu.Lock()
 					s.Status = "PENDING"
-					if s.Type == "push" {
-						s.LastCheck = time.Now()
-					}
 					if h, ok := e.GetHistory(s.ID); ok && len(h.Statuses) > 0 {
 						if h.Statuses[len(h.Statuses)-1] {
 							s.Status = "UP"
@@ -283,6 +346,9 @@ func (e *Engine) UpdateSiteConfig(site models.Site) {
 		site.LastCheck = existing.LastCheck
 		site.SentSSLWarning = existing.SentSSLWarning
 		site.FailureCount = existing.FailureCount
+		site.LastError = existing.LastError
+		site.StatusChangedAt = existing.StatusChangedAt
+		site.LastSuccessAt = existing.LastSuccessAt
 		e.liveState[site.ID] = site
 		e.addToTokenIndex(site)
 	}
@@ -393,33 +459,62 @@ func (e *Engine) checkByID(id int) {
 		updatedSite.CertExpiry = result.CertExpiry
 		updatedSite.Latency = time.Duration(result.LatencyNs)
 		updatedSite.LastCheck = time.Now()
-		e.handleStatusChange(updatedSite, result.Status, result.StatusCode, time.Duration(result.LatencyNs))
+		e.handleStatusChange(updatedSite, result.Status, result.StatusCode, time.Duration(result.LatencyNs), result.ErrorReason)
 	}
 }

 func (e *Engine) checkPush(site models.Site) {
-	deadline := site.LastCheck.Add(time.Duration(site.Interval) * time.Second).Add(pushGracePeriod)
-	if time.Now().After(deadline) {
-		e.handleStatusChange(site, "DOWN", 0, 0)
-	} else if site.Status != "UP" {
-		e.handleStatusChange(site, "UP", 200, 0)
+	if site.Status == "PENDING" {
+		return
+	}
+
+	interval := time.Duration(site.Interval) * time.Second
+	grace := interval / 2
+	if grace < minPushGrace {
+		grace = minPushGrace
+	}
+
+	overdue := site.LastCheck.Add(interval)
+	graceEnd := overdue.Add(grace)
+	now := time.Now()
+
+	if now.After(graceEnd) {
+		if site.Status != "DOWN" {
+			e.handleStatusChange(site, "DOWN", 0, 0, "heartbeat missed")
+		}
+	} else if now.After(overdue) {
+		if site.Status != "LATE" {
+			e.handleStatusChange(site, "LATE", 0, 0, "heartbeat overdue")
+		}
 	}
 }

-func (e *Engine) handleStatusChange(site models.Site, rawStatus string, code int, latency time.Duration) {
+func (e *Engine) handleStatusChange(site models.Site, rawStatus string, code int, latency time.Duration, errorReason string) {
 	if !e.IsActive() {
 		return
 	}

 	newState := site
 	newState.StatusCode = code
+	newState.LastError = errorReason
+
+	if rawStatus == "UP" {
+		newState.LastSuccessAt = time.Now()
+		newState.LastError = ""
+	} else {
+		newState.LastSuccessAt = site.LastSuccessAt
+	}

 	if site.Status == "UP" && rawStatus != "UP" {
 		newState.FailureCount++
 		if newState.FailureCount > site.MaxRetries {
 			newState.Status = rawStatus
 			newState.FailureCount = site.MaxRetries + 1
-			e.AddLog(fmt.Sprintf("Monitor '%s' confirmed DOWN", site.Name))
+			if errorReason != "" {
+				e.AddLog(fmt.Sprintf("Monitor '%s' confirmed DOWN: %s", site.Name, errorReason))
+			} else {
+				e.AddLog(fmt.Sprintf("Monitor '%s' confirmed DOWN", site.Name))
+			}
 		} else {
 			e.AddLog(fmt.Sprintf("Monitor '%s' failed check %d/%d", site.Name, newState.FailureCount, site.MaxRetries))
 		}
@@ -431,6 +526,14 @@ func (e *Engine) handleStatusChange(site models.Site, rawStatus string, code int
 		newState.FailureCount = site.MaxRetries + 1
 	}

+	if newState.Status != site.Status && site.Status != "PENDING" {
+		newState.StatusChangedAt = time.Now()
+	} else if site.StatusChangedAt.IsZero() && newState.Status != "PENDING" {
+		newState.StatusChangedAt = time.Now()
+	} else {
+		newState.StatusChangedAt = site.StatusChangedAt
+	}
+
 	inMaint := e.isInMaintenance(site.ID)

 	if site.Type == "http" && site.CheckSSL && site.HasSSL {
@@ -455,12 +558,24 @@ func (e *Engine) handleStatusChange(site models.Site, rawStatus string, code int

 	e.recordCheck(site.ID, latency, rawStatus == "UP")

+	if newState.Status != site.Status && site.Status != "PENDING" {
+		go func() { _ = e.db.SaveStateChange(site.ID, site.Status, newState.Status, errorReason) }()
+	}
+
 	isBroken := func(s string) bool { return s == "DOWN" || s == "SSL EXP" }
+
+	if site.Status == "UP" && newState.Status == "LATE" {
+		e.AddLog(fmt.Sprintf("Monitor '%s' heartbeat overdue", site.Name))
+	}
+
 	if !isBroken(site.Status) && isBroken(newState.Status) && newState.Status != "PENDING" {
 		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 errorReason != "" {
+				msg = fmt.Sprintf("Monitor '%s' is DOWN: %s", site.Name, errorReason)
+			}
 			if site.Type == "push" {
 				msg = fmt.Sprintf("Push Monitor '%s' missed heartbeat.", site.Name)
 			}
@@ -468,11 +583,17 @@ func (e *Engine) handleStatusChange(site models.Site, rawStatus string, code int
 		}
 	}
 	if isBroken(site.Status) && newState.Status == "UP" {
-		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))
+		downDur := ""
+		if !site.StatusChangedAt.IsZero() {
+			downDur = fmt.Sprintf(" (was down %s)", fmtDurationShort(time.Since(site.StatusChangedAt)))
 		}
+		e.AddLog(fmt.Sprintf("Monitor '%s' recovered%s", site.Name, downDur))
+		if !inMaint {
+			e.triggerAlert(site.AlertID, "✅ RECOVERY", fmt.Sprintf("Monitor '%s' is UP%s", site.Name, downDur))
+		}
+	}
+	if site.Status == "LATE" && newState.Status == "UP" && !isBroken(site.Status) {
+		e.AddLog(fmt.Sprintf("Monitor '%s' heartbeat arrived (was late)", site.Name))
 	}
 }

@@ -489,11 +610,69 @@ func (e *Engine) triggerAlert(alertID int, title, message string) {
 			defer cancel()
 			if err := provider.Send(ctx, title, message); err != nil {
 				e.AddLog(fmt.Sprintf("Alert send failed (%s): %v", cfg.Name, err))
+				e.recordAlertResult(alertID, false, err.Error())
+			} else {
+				e.recordAlertResult(alertID, true, "")
 			}
 		}()
 	}
 }

+func (e *Engine) recordAlertResult(alertID int, ok bool, errMsg string) {
+	e.alertHealthMu.Lock()
+	defer e.alertHealthMu.Unlock()
+	h := e.alertHealth[alertID]
+	h.LastSendAt = time.Now()
+	h.LastSendOK = ok
+	h.SendCount++
+	if ok {
+		h.LastError = ""
+	} else {
+		h.LastError = errMsg
+		h.FailCount++
+	}
+	e.alertHealth[alertID] = h
+
+	// Persist best-effort so health survives restarts; DB IO off the alert path.
+	go func(rec models.AlertHealthRecord) {
+		_ = e.db.SaveAlertHealth(rec)
+	}(models.AlertHealthRecord{
+		AlertID:    alertID,
+		LastSendAt: h.LastSendAt,
+		LastSendOK: h.LastSendOK,
+		LastError:  h.LastError,
+		SendCount:  h.SendCount,
+		FailCount:  h.FailCount,
+	})
+}
+
+func (e *Engine) GetAlertHealth(alertID int) AlertHealth {
+	e.alertHealthMu.RLock()
+	defer e.alertHealthMu.RUnlock()
+	return e.alertHealth[alertID]
+}
+
+func (e *Engine) TestAlert(alertID int) error {
+	cfg, err := e.db.GetAlert(alertID)
+	if err != nil {
+		return fmt.Errorf("failed to load alert: %w", err)
+	}
+	provider := alert.GetProvider(cfg)
+	if provider == nil {
+		return fmt.Errorf("no provider for type %q", cfg.Type)
+	}
+	ctx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
+	defer cancel()
+	err = provider.Send(ctx, "🧪 Test Alert", fmt.Sprintf("Test notification from uptop for channel '%s'.", cfg.Name))
+	if err != nil {
+		e.recordAlertResult(alertID, false, err.Error())
+		return err
+	}
+	e.recordAlertResult(alertID, true, "")
+	e.AddLog(fmt.Sprintf("Test alert sent to '%s'", cfg.Name))
+	return nil
+}
+
 func (e *Engine) isInMaintenance(monitorID int) bool {
 	inMaint, err := e.db.IsMonitorInMaintenance(monitorID)
 	if err != nil {
@@ -554,16 +733,17 @@ func (e *Engine) SetAggStrategy(strategy AggregationStrategy) {
 	e.aggStrategy = strategy
 }

-func (e *Engine) IngestProbeResult(nodeID string, siteID int, latencyNs int64, isUp bool) {
+func (e *Engine) IngestProbeResult(nodeID string, siteID int, latencyNs int64, isUp bool, errorReason string) {
 	e.probeResultsMu.Lock()
 	if e.probeResults[siteID] == nil {
 		e.probeResults[siteID] = make(map[string]NodeResult)
 	}
 	e.probeResults[siteID][nodeID] = NodeResult{
-		NodeID:    nodeID,
-		IsUp:      isUp,
-		LatencyNs: latencyNs,
-		CheckedAt: time.Now(),
+		NodeID:      nodeID,
+		IsUp:        isUp,
+		LatencyNs:   latencyNs,
+		CheckedAt:   time.Now(),
+		ErrorReason: errorReason,
 	}
 	results := make([]NodeResult, 0, len(e.probeResults[siteID]))
 	for _, r := range e.probeResults[siteID] {
@@ -588,7 +768,7 @@ func (e *Engine) IngestProbeResult(nodeID string, siteID int, latencyNs int64, i
 	updatedSite := site
 	updatedSite.Latency = time.Duration(avgLatency)
 	updatedSite.LastCheck = time.Now()
-	e.handleStatusChange(updatedSite, rawStatus, 0, time.Duration(avgLatency))
+	e.handleStatusChange(updatedSite, rawStatus, 0, time.Duration(avgLatency), errorReason)
 }

 func (e *Engine) GetProbeResults(siteID int) map[string]NodeResult {
@@ -601,3 +781,11 @@ func (e *Engine) GetProbeResults(siteID int) map[string]NodeResult {
 	}
 	return cp
 }
+
+func (e *Engine) GetStateChanges(siteID int, limit int) []models.StateChange {
+	changes, err := e.db.GetStateChanges(siteID, limit)
+	if err != nil {
+		return nil
+	}
+	return changes
+}
@@ -2,10 +2,11 @@ package monitor

 import (
 	"fmt"
-	"gitea.lerkolabs.com/lerko/uptop/internal/models"
 	"sync"
 	"testing"
 	"time"
+
+	"gitea.lerkolabs.com/lerko/uptop/internal/models"
 )

 // --- Mock Store ---
@@ -62,18 +63,24 @@ func (m *mockStore) GetNode(string) (models.ProbeNode, error)         { return m
 func (m *mockStore) GetAllNodes() ([]models.ProbeNode, error)         { return nil, nil }
 func (m *mockStore) UpdateNodeLastSeen(string) error                  { return nil }
 func (m *mockStore) DeleteNode(string) error                          { return nil }
+func (m *mockStore) LoadAlertHealth() (map[int]models.AlertHealthRecord, error) {
+	return nil, nil
+}
+func (m *mockStore) SaveAlertHealth(models.AlertHealthRecord) error { return 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) GetPreference(string) (string, error)                { return "", nil }
-func (m *mockStore) SetPreference(string, string) error                  { return nil }
-func (m *mockStore) Close() error                                        { return 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) GetPreference(string) (string, error)                   { return "", nil }
+func (m *mockStore) SetPreference(string, string) error                     { return nil }
+func (m *mockStore) SaveStateChange(int, string, string, string) error      { return nil }
+func (m *mockStore) GetStateChanges(int, int) ([]models.StateChange, error) { return nil, nil }
+func (m *mockStore) Close() error                                           { return nil }

 func (m *mockStore) GetAllAlerts() ([]models.AlertConfig, error) {
 	m.mu.Lock()
@@ -174,7 +181,7 @@ func TestHandleStatusChange_PendingToUp(t *testing.T) {
 	site := models.Site{ID: 1, Name: "test", Status: "PENDING", MaxRetries: 3, AlertID: 1}
 	injectSite(e, site)

-	e.handleStatusChange(site, "UP", 200, 10*time.Millisecond)
+	e.handleStatusChange(site, "UP", 200, 10*time.Millisecond, "")

 	s, _ := getSite(e, 1)
 	if s.Status != "UP" {
@@ -195,7 +202,7 @@ func TestHandleStatusChange_UpIncrementFailure(t *testing.T) {
 	site := models.Site{ID: 1, Name: "test", Status: "UP", MaxRetries: 3, FailureCount: 0}
 	injectSite(e, site)

-	e.handleStatusChange(site, "DOWN", 500, 0)
+	e.handleStatusChange(site, "DOWN", 500, 0, "test error")

 	s, _ := getSite(e, 1)
 	if s.Status != "UP" {
@@ -213,7 +220,7 @@ func TestHandleStatusChange_UpToDown_ExceedsRetries(t *testing.T) {
 	site := models.Site{ID: 1, Name: "test", Status: "UP", MaxRetries: 2, FailureCount: 2, AlertID: 1}
 	injectSite(e, site)

-	e.handleStatusChange(site, "DOWN", 500, 0)
+	e.handleStatusChange(site, "DOWN", 500, 0, "test error")

 	s, _ := getSite(e, 1)
 	if s.Status != "DOWN" {
@@ -236,7 +243,7 @@ func TestHandleStatusChange_UpToDown_ZeroRetries(t *testing.T) {
 	site := models.Site{ID: 1, Name: "test", Status: "UP", MaxRetries: 0, FailureCount: 0, AlertID: 1}
 	injectSite(e, site)

-	e.handleStatusChange(site, "DOWN", 0, 0)
+	e.handleStatusChange(site, "DOWN", 0, 0, "test error")

 	s, _ := getSite(e, 1)
 	if s.Status != "DOWN" {
@@ -255,7 +262,7 @@ func TestHandleStatusChange_DownToUp_Recovery(t *testing.T) {
 	site := models.Site{ID: 1, Name: "test", Status: "DOWN", FailureCount: 4, AlertID: 1}
 	injectSite(e, site)

-	e.handleStatusChange(site, "UP", 200, 5*time.Millisecond)
+	e.handleStatusChange(site, "UP", 200, 5*time.Millisecond, "")

 	s, _ := getSite(e, 1)
 	if s.Status != "UP" {
@@ -276,7 +283,7 @@ func TestHandleStatusChange_DownStaysDown(t *testing.T) {
 	site := models.Site{ID: 1, Name: "test", Status: "DOWN", MaxRetries: 2, FailureCount: 3}
 	injectSite(e, site)

-	e.handleStatusChange(site, "DOWN", 0, 0)
+	e.handleStatusChange(site, "DOWN", 0, 0, "test error")

 	s, _ := getSite(e, 1)
 	if s.Status != "DOWN" {
@@ -295,7 +302,7 @@ func TestHandleStatusChange_SSLExpired(t *testing.T) {
 	site := models.Site{ID: 1, Name: "test", Status: "UP", MaxRetries: 0, AlertID: 1}
 	injectSite(e, site)

-	e.handleStatusChange(site, "SSL EXP", 0, 0)
+	e.handleStatusChange(site, "SSL EXP", 0, 0, "SSL certificate expired")

 	s, _ := getSite(e, 1)
 	if s.Status != "SSL EXP" {
@@ -315,7 +322,7 @@ func TestHandleStatusChange_AlertSuppressedMaintenance(t *testing.T) {
 	site := models.Site{ID: 1, Name: "test", Status: "UP", MaxRetries: 0, AlertID: 1}
 	injectSite(e, site)

-	e.handleStatusChange(site, "DOWN", 0, 0)
+	e.handleStatusChange(site, "DOWN", 0, 0, "test error")

 	s, _ := getSite(e, 1)
 	if s.Status != "DOWN" {
@@ -346,7 +353,7 @@ func TestHandleStatusChange_RecoverySuppressedMaintenance(t *testing.T) {
 	site := models.Site{ID: 1, Name: "test", Status: "DOWN", AlertID: 1}
 	injectSite(e, site)

-	e.handleStatusChange(site, "UP", 200, 0)
+	e.handleStatusChange(site, "UP", 200, 0, "")

 	s, _ := getSite(e, 1)
 	if s.Status != "UP" {
@@ -370,7 +377,7 @@ func TestHandleStatusChange_SSLWarning(t *testing.T) {
 	}
 	injectSite(e, site)

-	e.handleStatusChange(site, "UP", 200, 0)
+	e.handleStatusChange(site, "UP", 200, 0, "")

 	s, _ := getSite(e, 1)
 	if !s.SentSSLWarning {
@@ -393,7 +400,7 @@ func TestHandleStatusChange_SSLWarningNotRepeated(t *testing.T) {
 	}
 	injectSite(e, site)

-	e.handleStatusChange(site, "UP", 200, 0)
+	e.handleStatusChange(site, "UP", 200, 0, "")

 	waitAsync()
 	if len(ms.getAlertCallsSnapshot()) != 0 {
@@ -412,7 +419,7 @@ func TestHandleStatusChange_SSLWarningReset(t *testing.T) {
 	}
 	injectSite(e, site)

-	e.handleStatusChange(site, "UP", 200, 0)
+	e.handleStatusChange(site, "UP", 200, 0, "")

 	s, _ := getSite(e, 1)
 	if s.SentSSLWarning {
@@ -433,7 +440,7 @@ func TestHandleStatusChange_SSLWarningSuppressedMaint(t *testing.T) {
 	}
 	injectSite(e, site)

-	e.handleStatusChange(site, "UP", 200, 0)
+	e.handleStatusChange(site, "UP", 200, 0, "")

 	s, _ := getSite(e, 1)
 	if !s.SentSSLWarning {
@@ -452,7 +459,7 @@ func TestHandleStatusChange_InactiveEngine(t *testing.T) {
 	injectSite(e, site)
 	e.SetActive(false)

-	e.handleStatusChange(site, "DOWN", 0, 0)
+	e.handleStatusChange(site, "DOWN", 0, 0, "test error")

 	s, _ := getSite(e, 1)
 	if s.Status != "UP" {
@@ -534,7 +541,7 @@ func TestCheckPush_DeadlineMissed(t *testing.T) {
 	site := models.Site{
 		ID: 1, Name: "push", Type: "push", Status: "UP",
 		Interval: 10, MaxRetries: 0,
-		LastCheck: time.Now().Add(-20 * time.Second),
+		LastCheck: time.Now().Add(-120 * time.Second),
 	}
 	injectSite(e, site)

@@ -546,6 +553,24 @@ func TestCheckPush_DeadlineMissed(t *testing.T) {
 	}
 }

+func TestCheckPush_OverdueBecomesLate(t *testing.T) {
+	ms := newMockStore()
+	e := newTestEngine(ms)
+	site := models.Site{
+		ID: 1, Name: "push", Type: "push", Status: "UP",
+		Interval:  300,
+		LastCheck: time.Now().Add(-310 * time.Second),
+	}
+	injectSite(e, site)
+
+	e.checkPush(site)
+
+	s, _ := getSite(e, 1)
+	if s.Status != "LATE" {
+		t.Errorf("expected LATE when overdue but within grace, got %s", s.Status)
+	}
+}
+
 func TestCheckPush_WithinDeadline(t *testing.T) {
 	ms := newMockStore()
 	e := newTestEngine(ms)
@@ -563,20 +588,20 @@ func TestCheckPush_WithinDeadline(t *testing.T) {
 	}
 }

-func TestCheckPush_PendingToUp(t *testing.T) {
+func TestCheckPush_PendingStaysPending(t *testing.T) {
 	ms := newMockStore()
 	e := newTestEngine(ms)
 	site := models.Site{
 		ID: 1, Name: "push", Type: "push", Status: "PENDING",
-		Interval: 60, LastCheck: time.Now(),
+		Interval: 60,
 	}
 	injectSite(e, site)

 	e.checkPush(site)

 	s, _ := getSite(e, 1)
-	if s.Status != "UP" {
-		t.Errorf("expected UP, got %s", s.Status)
+	if s.Status != "PENDING" {
+		t.Errorf("expected PENDING to stay until first heartbeat, got %s", s.Status)
 	}
 }

@@ -991,7 +1016,7 @@ func TestConcurrent_HandleStatusChangeAndGetState(t *testing.T) {
 		wg.Add(2)
 		go func() {
 			defer wg.Done()
-			e.handleStatusChange(site, "DOWN", 500, 0)
+			e.handleStatusChange(site, "DOWN", 500, 0, "test error")
 		}()
 		go func() {
 			defer wg.Done()
@@ -67,6 +67,7 @@ var statusTpl = template.Must(template.New("status").Parse(`
 		.UP { background: #9ece6a; color: #1a1b26; }
 		.DOWN { background: #f7768e; color: #1a1b26; }
 		.PENDING { background: #e0af68; color: #1a1b26; }
+		.LATE { background: #e0af68; color: #1a1b26; }
 		.SSL-EXP { background: #e0af68; color: #1a1b26; }
 		.PAUSED { background: #565f89; color: #c0caf5; }
 		.MAINT { background: #bb9af7; color: #1a1b26; }
@@ -403,9 +404,10 @@ func Start(cfg ServerConfig, s store.Store, eng *monitor.Engine) *http.Server {
 		var req struct {
 			NodeID  string `json:"node_id"`
 			Results []struct {
-				SiteID    int   `json:"site_id"`
-				LatencyNs int64 `json:"latency_ns"`
-				IsUp      bool  `json:"is_up"`
+				SiteID      int    `json:"site_id"`
+				LatencyNs   int64  `json:"latency_ns"`
+				IsUp        bool   `json:"is_up"`
+				ErrorReason string `json:"error_reason"`
 			} `json:"results"`
 		}
 		if err := json.NewDecoder(r.Body).Decode(&req); err != nil {
@@ -420,7 +422,7 @@ func Start(cfg ServerConfig, s store.Store, eng *monitor.Engine) *http.Server {
 			if err := s.SaveCheckFromNode(result.SiteID, req.NodeID, result.LatencyNs, result.IsUp); err != nil {
 				log.Printf("Failed to save probe result: %v", err)
 			}
-			eng.IngestProbeResult(req.NodeID, result.SiteID, result.LatencyNs, result.IsUp)
+			eng.IngestProbeResult(req.NodeID, result.SiteID, result.LatencyNs, result.IsUp, result.ErrorReason)
 		}
 		if err := s.UpdateNodeLastSeen(req.NodeID); err != nil {
 			log.Printf("Failed to update node last seen: %v", err)
@@ -4,13 +4,14 @@ import (
 	"bytes"
 	"encoding/json"
 	"fmt"
-	"gitea.lerkolabs.com/lerko/uptop/internal/models"
-	"gitea.lerkolabs.com/lerko/uptop/internal/monitor"
 	"net"
 	"net/http"
 	"sync"
 	"testing"
 	"time"
+
+	"gitea.lerkolabs.com/lerko/uptop/internal/models"
+	"gitea.lerkolabs.com/lerko/uptop/internal/monitor"
 )

 // --- Mock Store ---
@@ -64,18 +65,24 @@ func (m *mockStore) AddAlertReturningID(string, string, map[string]string) (int,
 func (m *mockStore) GetAllNodes() ([]models.ProbeNode, error) { return nil, nil }
 func (m *mockStore) UpdateNodeLastSeen(string) error          { return nil }
 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) LoadAlertHealth() (map[int]models.AlertHealthRecord, error) {
+	return nil, nil
+}
+func (m *mockStore) SaveAlertHealth(models.AlertHealthRecord) error { return nil }
+func (m *mockStore) SaveLog(string) error                           { return nil }
+func (m *mockStore) LoadLogs(int) ([]string, 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 (m *mockStore) GetPreference(string) (string, error)                { return "", nil }
-func (m *mockStore) SetPreference(string, string) error                  { return nil }
-func (m *mockStore) Close() error                                        { return 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 (m *mockStore) GetPreference(string) (string, error)                   { return "", nil }
+func (m *mockStore) SetPreference(string, string) error                     { return nil }
+func (m *mockStore) SaveStateChange(int, string, string, string) error      { return nil }
+func (m *mockStore) GetStateChanges(int, int) ([]models.StateChange, error) { return nil, nil }
+func (m *mockStore) Close() error                                           { return nil }

 func (m *mockStore) ExportData() (models.Backup, error) {
 	return models.Backup{
@@ -14,6 +14,7 @@ type Dialect interface {
 	ImportWipe(tx *sql.Tx)
 	ImportResetSequences(tx *sql.Tx)
 	UpsertNodeSQL() string
+	UpsertAlertHealthSQL() string
 }

 func rewritePlaceholders(query string, dollarStyle bool) string {
@@ -72,6 +72,23 @@ func (d *PostgresDialect) CreateTablesSQL() []string {
 			key TEXT PRIMARY KEY,
 			value TEXT NOT NULL
 		)`,
+		`CREATE TABLE IF NOT EXISTS state_changes (
+			id SERIAL PRIMARY KEY,
+			site_id INTEGER NOT NULL,
+			from_status TEXT NOT NULL,
+			to_status TEXT NOT NULL,
+			error_reason TEXT DEFAULT '',
+			changed_at TIMESTAMP DEFAULT NOW()
+		)`,
+		`CREATE INDEX IF NOT EXISTS idx_state_changes_site ON state_changes(site_id, changed_at DESC)`,
+		`CREATE TABLE IF NOT EXISTS alert_health (
+			alert_id INTEGER PRIMARY KEY,
+			last_send_at TIMESTAMP,
+			last_send_ok BOOLEAN DEFAULT FALSE,
+			last_error TEXT DEFAULT '',
+			send_count INTEGER DEFAULT 0,
+			fail_count INTEGER DEFAULT 0
+		)`,
 	}
 }

@@ -97,6 +114,10 @@ func (d *PostgresDialect) UpsertNodeSQL() string {
 	return "INSERT INTO nodes (id, name, region, last_seen, version) VALUES ($1, $2, $3, NOW(), $4) ON CONFLICT (id) DO UPDATE SET name = EXCLUDED.name, region = EXCLUDED.region, last_seen = NOW(), version = EXCLUDED.version"
 }

+func (d *PostgresDialect) UpsertAlertHealthSQL() string {
+	return "INSERT INTO alert_health (alert_id, last_send_at, last_send_ok, last_error, send_count, fail_count) VALUES ($1, $2, $3, $4, $5, $6) ON CONFLICT (alert_id) DO UPDATE SET last_send_at = EXCLUDED.last_send_at, last_send_ok = EXCLUDED.last_send_ok, last_error = EXCLUDED.last_error, send_count = EXCLUDED.send_count, fail_count = EXCLUDED.fail_count"
+}
+
 func (d *PostgresDialect) ResetSequenceOnEmpty(db *sql.DB, table string) {}

 func (d *PostgresDialect) ImportWipe(tx *sql.Tx) {
@@ -79,6 +79,23 @@ func (d *SQLiteDialect) CreateTablesSQL() []string {
 			key TEXT PRIMARY KEY,
 			value TEXT NOT NULL
 		)`,
+		`CREATE TABLE IF NOT EXISTS state_changes (
+			id INTEGER PRIMARY KEY AUTOINCREMENT,
+			site_id INTEGER NOT NULL,
+			from_status TEXT NOT NULL,
+			to_status TEXT NOT NULL,
+			error_reason TEXT DEFAULT '',
+			changed_at DATETIME DEFAULT CURRENT_TIMESTAMP
+		)`,
+		`CREATE INDEX IF NOT EXISTS idx_state_changes_site ON state_changes(site_id, changed_at DESC)`,
+		`CREATE TABLE IF NOT EXISTS alert_health (
+			alert_id INTEGER PRIMARY KEY,
+			last_send_at DATETIME,
+			last_send_ok BOOLEAN DEFAULT 0,
+			last_error TEXT DEFAULT '',
+			send_count INTEGER DEFAULT 0,
+			fail_count INTEGER DEFAULT 0
+		)`,
 	}
 }

@@ -104,6 +121,10 @@ func (d *SQLiteDialect) UpsertNodeSQL() string {
 	return "INSERT OR REPLACE INTO nodes (id, name, region, last_seen, version) VALUES (?, ?, ?, CURRENT_TIMESTAMP, ?)"
 }

+func (d *SQLiteDialect) UpsertAlertHealthSQL() string {
+	return "INSERT OR REPLACE INTO alert_health (alert_id, last_send_at, last_send_ok, last_error, send_count, fail_count) VALUES (?, ?, ?, ?, ?, ?)"
+}
+
 func (d *SQLiteDialect) ResetSequenceOnEmpty(db *sql.DB, table string) {
 	var count int
 	_ = db.QueryRow("SELECT COUNT(*) FROM " + table).Scan(&count) //nolint:errcheck
@@ -347,6 +347,29 @@ func (s *SQLStore) DeleteUser(id int) error {
 	return err
 }

+func (s *SQLStore) SaveStateChange(siteID int, fromStatus, toStatus, errorReason string) error {
+	_, err := s.db.Exec(s.q("INSERT INTO state_changes (site_id, from_status, to_status, error_reason) VALUES (?, ?, ?, ?)"),
+		siteID, fromStatus, toStatus, errorReason)
+	return err
+}
+
+func (s *SQLStore) GetStateChanges(siteID int, limit int) ([]models.StateChange, error) {
+	rows, err := s.db.Query(s.q("SELECT id, site_id, from_status, to_status, error_reason, changed_at FROM state_changes WHERE site_id = ? ORDER BY changed_at DESC LIMIT ?"), siteID, limit)
+	if err != nil {
+		return nil, err
+	}
+	defer rows.Close()
+	var changes []models.StateChange
+	for rows.Next() {
+		var sc models.StateChange
+		if err := rows.Scan(&sc.ID, &sc.SiteID, &sc.FromStatus, &sc.ToStatus, &sc.ErrorReason, &sc.ChangedAt); err != nil {
+			return changes, err
+		}
+		changes = append(changes, sc)
+	}
+	return changes, rows.Err()
+}
+
 func (s *SQLStore) SaveCheck(siteID int, latencyNs int64, isUp bool) error {
 	return s.SaveCheckFromNode(siteID, "", latencyNs, isUp)
 }
@@ -407,6 +430,37 @@ func (s *SQLStore) DeleteNode(id string) error {
 	return err
 }

+func (s *SQLStore) LoadAlertHealth() (map[int]models.AlertHealthRecord, error) {
+	rows, err := s.db.Query("SELECT alert_id, last_send_at, last_send_ok, last_error, send_count, fail_count FROM alert_health")
+	if err != nil {
+		return nil, err
+	}
+	defer rows.Close()
+	out := make(map[int]models.AlertHealthRecord)
+	for rows.Next() {
+		var r models.AlertHealthRecord
+		var lastSend sql.NullTime
+		if err := rows.Scan(&r.AlertID, &lastSend, &r.LastSendOK, &r.LastError, &r.SendCount, &r.FailCount); err != nil {
+			return out, err
+		}
+		if lastSend.Valid {
+			r.LastSendAt = lastSend.Time
+		}
+		out[r.AlertID] = r
+	}
+	return out, rows.Err()
+}
+
+func (s *SQLStore) SaveAlertHealth(h models.AlertHealthRecord) error {
+	var lastSend interface{}
+	if !h.LastSendAt.IsZero() {
+		lastSend = h.LastSendAt
+	}
+	_, err := s.db.Exec(s.dialect.UpsertAlertHealthSQL(),
+		h.AlertID, lastSend, h.LastSendOK, h.LastError, h.SendCount, h.FailCount)
+	return err
+}
+
 func (s *SQLStore) SaveLog(message string) error {
 	_, err := s.db.Exec(s.q("INSERT INTO logs (message) VALUES (?)"), message)
 	if err != nil {
@@ -38,6 +38,10 @@ type Store interface {
 	SaveCheckFromNode(siteID int, nodeID string, latencyNs int64, isUp bool) error
 	LoadAllHistory(limit int) (map[int][]models.CheckRecord, error)

+	// State Changes
+	SaveStateChange(siteID int, fromStatus, toStatus, errorReason string) error
+	GetStateChanges(siteID int, limit int) ([]models.StateChange, error)
+
 	// Nodes
 	RegisterNode(node models.ProbeNode) error
 	GetNode(id string) (models.ProbeNode, error)
@@ -45,6 +49,10 @@ type Store interface {
 	UpdateNodeLastSeen(id string) error
 	DeleteNode(id string) error

+	// Alert Health
+	LoadAlertHealth() (map[int]models.AlertHealthRecord, error)
+	SaveAlertHealth(h models.AlertHealthRecord) error
+
 	// Logs
 	SaveLog(message string) error
 	LoadLogs(limit int) ([]string, error)
@@ -2,7 +2,10 @@ package tui

 import (
 	"fmt"
+	"strings"
+	"time"

+	"gitea.lerkolabs.com/lerko/uptop/internal/monitor"
 	tea "github.com/charmbracelet/bubbletea"
 	"github.com/charmbracelet/huh"
 	"github.com/charmbracelet/lipgloss"
@@ -113,34 +116,122 @@ func fmtAlertConfig(alert struct {
 	}
 }

+func fmtAlertHealth(h monitor.AlertHealth) string {
+	if h.LastSendAt.IsZero() {
+		return subtleStyle.Render("●")
+	}
+	if h.LastSendOK {
+		return specialStyle.Render("●")
+	}
+	return dangerStyle.Render("●")
+}
+
+func fmtAlertLastSent(h monitor.AlertHealth) string {
+	if h.LastSendAt.IsZero() {
+		return subtleStyle.Render("never")
+	}
+	d := time.Since(h.LastSendAt)
+	if d < time.Minute {
+		return fmt.Sprintf("%ds ago", int(d.Seconds()))
+	}
+	if d < time.Hour {
+		return fmt.Sprintf("%dm ago", int(d.Minutes()))
+	}
+	if d < 24*time.Hour {
+		return fmt.Sprintf("%dh ago", int(d.Hours()))
+	}
+	return fmt.Sprintf("%dd ago", int(d.Hours())/24)
+}
+
 func (m Model) viewAlertsTab() string {
 	if len(m.alerts) == 0 {
 		return "\n  No alert channels configured. Press [n] to add one."
 	}

+	var headers []string
+	var widths []int
+	if m.isWide() {
+		headers = []string{"#", "", "NAME", "TYPE", "CONFIG", "LAST SENT"}
+		widths = []int{4, 3, 18, 12, 40, 12}
+	} else {
+		headers = []string{"#", "", "NAME", "TYPE", "CONFIG", "SENT"}
+		widths = []int{4, 3, 14, 10, 24, 8}
+	}
+	nameW := widths[2]
+	cfgW := widths[4]
+
 	return m.renderTable(
-		[]string{"#", "NAME", "TYPE", "CONFIG"},
+		headers,
 		len(m.alerts),
 		func(start, end int) [][]string {
 			var rows [][]string
 			for i := start; i < end; i++ {
 				a := m.alerts[i]
+				h := m.engine.GetAlertHealth(a.ID)
 				rows = append(rows, []string{
 					fmt.Sprintf("%d", i+1),
-					m.zones.Mark(fmt.Sprintf("alert-%d", i), limitStr(a.Name, 15)),
+					fmtAlertHealth(h),
+					m.zones.Mark(fmt.Sprintf("alert-%d", i), limitStr(a.Name, nameW-2)),
 					fmtAlertType(a.Type),
-					fmtAlertConfig(struct {
+					limitStr(fmtAlertConfig(struct {
 						Type     string
 						Settings map[string]string
-					}{a.Type, a.Settings}),
+					}{a.Type, a.Settings}), cfgW-2),
+					fmtAlertLastSent(h),
 				})
 			}
 			return rows
 		},
-		nil, nil,
+		widths, nil,
 	)
 }

+func (m Model) viewAlertDetailPanel() string {
+	if m.cursor >= len(m.alerts) {
+		return ""
+	}
+	a := m.alerts[m.cursor]
+	h := m.engine.GetAlertHealth(a.ID)
+
+	var b strings.Builder
+
+	b.WriteString(subtleStyle.Render("  Alerts > ") + titleStyle.Render(a.Name) + "\n\n")
+
+	row := func(label, value string) {
+		fmt.Fprintf(&b, "  %-16s %s\n", subtleStyle.Render(label), value)
+	}
+
+	row("Type", fmtAlertType(a.Type))
+
+	if h.LastSendAt.IsZero() {
+		row("Health", subtleStyle.Render("never sent"))
+	} else if h.LastSendOK {
+		row("Health", specialStyle.Render("OK"))
+	} else {
+		row("Health", dangerStyle.Render("FAILED"))
+	}
+
+	if !h.LastSendAt.IsZero() {
+		row("Last Sent", h.LastSendAt.Format("2006-01-02 15:04:05")+" ("+fmtAlertLastSent(h)+")")
+	}
+	if h.SendCount > 0 {
+		row("Sends", fmt.Sprintf("%d sent, %d failed", h.SendCount, h.FailCount))
+	}
+	if h.LastError != "" {
+		row("Last Error", dangerStyle.Render(limitStr(h.LastError, 60)))
+	}
+
+	b.WriteString("\n" + subtleStyle.Render("  CONFIGURATION") + "\n")
+	for k, v := range a.Settings {
+		row(k, v)
+	}
+
+	b.WriteString("\n\n")
+	b.WriteString(subtleStyle.Render("  [i/Esc] Back  [e] Edit  [t] Test  [q] Quit"))
+
+	return lipgloss.NewStyle().Padding(1, 2).Render(b.String())
+}
+
 func (m *Model) initAlertHuhForm() tea.Cmd {
 	m.alertFormData = &alertFormData{
 		AlertType:         "discord",
@@ -5,27 +5,83 @@ import (
 	"strings"
 )

-func colorizeLog(line string) string {
+type logSeverity int
+
+const (
+	severityInfo logSeverity = iota
+	severityWarn
+	severityDown
+	severityUp
+	severitySystem
+)
+
+func classifyLog(line string) logSeverity {
 	lower := strings.ToLower(line)
 	switch {
 	case strings.Contains(lower, "confirmed down"),
 		strings.Contains(lower, "is down"),
 		strings.Contains(lower, "missed heartbeat"),
-		strings.Contains(lower, "failed check"),
-		strings.Contains(lower, "ssl warning"):
-		return dangerStyle.Render(line)
+		strings.Contains(lower, "alert send failed"):
+		return severityDown
 	case strings.Contains(lower, "recovered"),
 		strings.Contains(lower, "is up"),
-		strings.Contains(lower, "recovery"):
-		return specialStyle.Render(line)
+		strings.Contains(lower, "recovery"),
+		strings.Contains(lower, "first heartbeat"):
+		return severityUp
+	case strings.Contains(lower, "failed check"),
+		strings.Contains(lower, "ssl warning"),
+		strings.Contains(lower, "overdue"),
+		strings.Contains(lower, "was late"):
+		return severityWarn
 	case strings.Contains(lower, "engine"),
-		strings.Contains(lower, "cluster"):
-		return titleStyle.Render(line)
+		strings.Contains(lower, "cluster"),
+		strings.Contains(lower, "loaded"),
+		strings.Contains(lower, "paused"),
+		strings.Contains(lower, "resumed"):
+		return severitySystem
 	default:
-		return line
+		return severityInfo
 	}
 }

+func isImportantLog(sev logSeverity) bool {
+	return sev == severityDown || sev == severityUp || sev == severitySystem
+}
+
+func renderLogTag(sev logSeverity) string {
+	switch sev {
+	case severityDown:
+		return dangerStyle.Render(" DOWN ")
+	case severityUp:
+		return specialStyle.Render("  UP  ")
+	case severityWarn:
+		return warnStyle.Render(" WARN ")
+	case severitySystem:
+		return titleStyle.Render(" SYS  ")
+	default:
+		return subtleStyle.Render(" info ")
+	}
+}
+
+func renderLogLine(line string) string {
+	sev := classifyLog(line)
+	tag := renderLogTag(sev)
+
+	ts := ""
+	msg := line
+	if len(line) > 10 && line[0] == '[' {
+		if idx := strings.Index(line, "]"); idx > 0 && idx < 12 {
+			ts = subtleStyle.Render(line[1:idx])
+			msg = strings.TrimSpace(line[idx+1:])
+		}
+	}
+
+	if ts != "" {
+		return fmt.Sprintf("  %s  %s  %s", ts, tag, msg)
+	}
+	return fmt.Sprintf("  %s  %s", tag, msg)
+}
+
 func (m Model) viewLogsTab() string {
 	content := m.logViewport.View()
 	if strings.TrimSpace(content) == "" || content == "Waiting for logs..." {
@@ -33,22 +89,34 @@ func (m Model) viewLogsTab() string {
 	}

 	lines := strings.Split(content, "\n")
-	var colored []string
+	var rendered []string
+	total := 0
+	shown := 0
+
 	for _, line := range lines {
-		if line == "" {
-			colored = append(colored, line)
+		if strings.TrimSpace(line) == "" {
 			continue
 		}
-		colored = append(colored, colorizeLog(line))
-	}
-
-	count := 0
-	for _, l := range lines {
-		if strings.TrimSpace(l) != "" {
-			count++
+		total++
+		sev := classifyLog(line)
+		if m.logFilterImportant && !isImportantLog(sev) {
+			continue
 		}
+		shown++
+		rendered = append(rendered, renderLogLine(line))
 	}

-	header := subtleStyle.Render(fmt.Sprintf("  %d entries  [↑/↓] Scroll  [PgUp/PgDn] Page", count))
-	return "\n" + header + "\n\n" + strings.Join(colored, "\n")
+	filterLabel := "All"
+	if m.logFilterImportant {
+		filterLabel = "Important"
+	}
+
+	header := subtleStyle.Render(fmt.Sprintf(
+		"  %d entries  [↑/↓] Scroll  [PgUp/PgDn] Page  [f] Filter: %s", shown, filterLabel))
+
+	if m.logFilterImportant && shown < total {
+		header += subtleStyle.Render(fmt.Sprintf("  (%d hidden)", total-shown))
+	}
+
+	return "\n" + header + "\n\n" + strings.Join(rendered, "\n")
 }
@@ -2,10 +2,11 @@ package tui

 import (
 	"fmt"
-	"gitea.lerkolabs.com/lerko/uptop/internal/models"
 	"strconv"
 	"time"

+	"gitea.lerkolabs.com/lerko/uptop/internal/models"
+
 	tea "github.com/charmbracelet/bubbletea"
 	"github.com/charmbracelet/huh"
 	"github.com/charmbracelet/lipgloss"
@@ -40,19 +41,19 @@ func fmtMaintType(t string) string {
 	return maintStyle.Render("maintenance")
 }

-func fmtMaintMonitor(monitorID int, sites []models.Site) string {
+func fmtMaintMonitorW(monitorID int, sites []models.Site, maxW int) string {
 	if monitorID == 0 {
 		return "All"
 	}
 	for _, s := range sites {
 		if s.ID == monitorID {
-			return limitStr(s.Name, 18)
+			return limitStr(s.Name, maxW)
 		}
 	}
 	return fmt.Sprintf("#%d", monitorID)
 }

-func fmtMaintTime(t time.Time) string {
+func fmtMaintTime(t time.Time, colW int) string {
 	if t.IsZero() {
 		return subtleStyle.Render("—")
 	}
@@ -60,7 +61,10 @@ func fmtMaintTime(t time.Time) string {
 	if t.Year() == now.Year() && t.YearDay() == now.YearDay() {
 		return t.Format("15:04")
 	}
-	return t.Format("15:04 Jan 02")
+	if colW >= 14 {
+		return t.Format("15:04 Jan 02")
+	}
+	return t.Format("Jan 02")
 }

 func (m Model) isMonitorInMaintenance(monitorID int) bool {
@@ -92,8 +96,21 @@ func (m Model) viewMaintTab() string {
 		return "\n  No maintenance windows or incidents. Press [n] to create one."
 	}

+	var headers []string
+	var widths []int
+	if m.isWide() {
+		headers = []string{"#", "TITLE", "TYPE", "MONITORS", "STATUS", "STARTED", "ENDS"}
+		widths = []int{4, 24, 14, 22, 12, 16, 16}
+	} else {
+		headers = []string{"#", "TITLE", "TYPE", "MON", "ST", "START", "ENDS"}
+		widths = []int{4, 14, 13, 14, 11, 14, 14}
+	}
+	titleW := widths[1]
+	monW := widths[3]
+	timeW := widths[5]
+
 	return m.renderTable(
-		[]string{"#", "TITLE", "TYPE", "MONITORS", "STATUS", "STARTED", "ENDS"},
+		headers,
 		len(m.maintenanceWindows),
 		func(start, end int) [][]string {
 			var rows [][]string
@@ -102,17 +119,17 @@ func (m Model) viewMaintTab() string {
 				mw := m.maintenanceWindows[i]
 				rows = append(rows, []string{
 					strconv.Itoa(i + 1),
-					m.zones.Mark(fmt.Sprintf("maint-%d", i), limitStr(mw.Title, 24)),
+					m.zones.Mark(fmt.Sprintf("maint-%d", i), limitStr(mw.Title, titleW-2)),
 					fmtMaintType(mw.Type),
-					fmtMaintMonitor(mw.MonitorID, allSites),
+					fmtMaintMonitorW(mw.MonitorID, allSites, monW-2),
 					fmtMaintStatus(mw),
-					fmtMaintTime(mw.StartTime),
-					fmtMaintTime(mw.EndTime),
+					fmtMaintTime(mw.StartTime, timeW),
+					fmtMaintTime(mw.EndTime, timeW),
 				})
 			}
 			return rows
 		},
-		[]int{6, 0, 14, 20, 12, 16, 16},
+		widths,
 		nil,
 	)
 }
@@ -10,16 +10,25 @@ func (m Model) viewNodesTab() string {
 		return "\n  No probe nodes connected."
 	}

-	colWidths := []int{0, 12, 20, 10, 8}
+	var headers []string
+	var widths []int
+	if m.isWide() {
+		headers = []string{"NAME", "REGION", "LAST SEEN", "VERSION", "STATUS"}
+		widths = []int{24, 14, 16, 12, 10}
+	} else {
+		headers = []string{"NAME", "REGION", "SEEN", "VER", "STATUS"}
+		widths = []int{16, 10, 10, 8, 8}
+	}
+	nameW := widths[0]

 	return m.renderTable(
-		[]string{"NAME", "REGION", "LAST SEEN", "VERSION", "STATUS"},
+		headers,
 		len(m.nodes),
 		func(start, end int) [][]string {
 			var rows [][]string
 			for i := start; i < end; i++ {
 				node := m.nodes[i]
-				name := limitStr(node.Name, 20)
+				name := limitStr(node.Name, nameW-2)
 				if name == "" {
 					name = node.ID
 				}
@@ -37,7 +46,7 @@ func (m Model) viewNodesTab() string {
 			}
 			return rows
 		},
-		colWidths,
+		widths,
 		nil,
 	)
 }
@@ -60,14 +60,18 @@ type siteFormData struct {
 	Regions       string
 }

-func latencySparkline(latencies []time.Duration, width int) string {
+func latencySparkline(latencies []time.Duration, statuses []bool, width int) string {
 	if len(latencies) == 0 {
 		return subtleStyle.Render(strings.Repeat("·", width))
 	}

 	samples := latencies
+	sampledStatuses := statuses
 	if len(samples) > width {
 		samples = samples[len(samples)-width:]
+		if len(sampledStatuses) > width {
+			sampledStatuses = sampledStatuses[len(sampledStatuses)-width:]
+		}
 	}

 	minL, maxL := samples[0], samples[0]
@@ -85,7 +89,7 @@ func latencySparkline(latencies []time.Duration, width int) string {
 		sb.WriteString(subtleStyle.Render(strings.Repeat("·", remaining)))
 	}
 	spread := maxL - minL
-	for _, l := range samples {
+	for i, l := range samples {
 		idx := 0
 		if spread > 0 {
 			idx = int(float64(l-minL) / float64(spread) * 7)
@@ -94,13 +98,18 @@ func latencySparkline(latencies []time.Duration, width int) string {
 			}
 		}
 		ch := string(sparkChars[idx])
-		ms := l.Milliseconds()
-		if ms < 200 {
-			sb.WriteString(specialStyle.Render(ch))
-		} else if ms < 500 {
-			sb.WriteString(warnStyle.Render(ch))
-		} else {
+		isDown := i < len(sampledStatuses) && !sampledStatuses[i]
+		if isDown {
 			sb.WriteString(dangerStyle.Render(ch))
+		} else {
+			ms := l.Milliseconds()
+			if ms < 200 {
+				sb.WriteString(specialStyle.Render(ch))
+			} else if ms < 500 {
+				sb.WriteString(warnStyle.Render(ch))
+			} else {
+				sb.WriteString(dangerStyle.Render(ch))
+			}
 		}
 	}
 	return sb.String()
@@ -302,6 +311,8 @@ func fmtStatus(status string, paused bool, inMaint bool) string {
 	switch status {
 	case "DOWN", "SSL EXP":
 		return dangerStyle.Render(status)
+	case "LATE":
+		return warnStyle.Render(status)
 	case "PENDING":
 		return subtleStyle.Render(status)
 	default:
@@ -309,28 +320,94 @@ func fmtStatus(status string, paused bool, inMaint bool) string {
 	}
 }

-func (m Model) dynamicWidths() (nameW, sparkW int) {
-	fixed := 6 + 10 + 10 + 8 + 8 + 7 + 9 // #, TYPE, STATUS, LATENCY, UPTIME, SSL, RETRY
-	overhead := 30                       // cell padding + borders
-	avail := m.termWidth - chromePadH - 2 - fixed - overhead
-	if avail < 30 {
-		avail = 30
+func fmtDuration(d time.Duration) string {
+	if d < time.Minute {
+		return fmt.Sprintf("%ds", int(d.Seconds()))
+	}
+	if d < time.Hour {
+		return fmt.Sprintf("%dm", int(d.Minutes()))
+	}
+	if d < 24*time.Hour {
+		h := int(d.Hours())
+		m := int(d.Minutes()) % 60
+		if m > 0 {
+			return fmt.Sprintf("%dh %dm", h, m)
+		}
+		return fmt.Sprintf("%dh", h)
+	}
+	days := int(d.Hours()) / 24
+	hours := int(d.Hours()) % 24
+	if hours > 0 {
+		return fmt.Sprintf("%dd %dh", days, hours)
+	}
+	return fmt.Sprintf("%dd", days)
+}
+
+type tableLayout struct {
+	nameW, sparkW int
+	headers       []string
+	colWidths     []int
+}
+
+func (m Model) computeLayout() tableLayout {
+	wide := m.isWide()
+
+	var fixed int
+	var headers []string
+	var widths []int
+
+	if wide {
+		//              #   NAME TYPE   STATUS LATENCY UPTIME HISTORY SSL RETRIES
+		headers = []string{"#", "NAME", "TYPE", "STATUS", "LATENCY", "UPTIME", "HISTORY", "SSL", "RETRIES"}
+		widths = []int{4, 0, 10, 10, 10, 8, 0, 7, 9}
+		fixed = 4 + 10 + 10 + 10 + 8 + 7 + 9
+	} else {
+		//              #   NAME TYPE   STATUS LAT   UP%   HISTORY SSL RT
+		headers = []string{"#", "NAME", "TYPE", "STATUS", "LAT", "UP%", "HISTORY", "SSL", "RT"}
+		widths = []int{4, 0, 8, 8, 7, 8, 0, 5, 5}
+		fixed = 4 + 8 + 8 + 7 + 8 + 5 + 5
+	}
+
+	numCols := len(headers)
+	borderOverhead := 2 + (numCols - 1)
+	avail := m.termWidth - chromePadH - 2 - borderOverhead - fixed
+	if avail < 20 {
+		avail = 20
+	}
+
+	maxName := 0
+	for _, s := range m.sites {
+		if n := len([]rune(s.Name)); n > maxName {
+			maxName = n
+		}
+	}
+	maxName += 4
+
+	nameW := avail / 2
+	if nameW > maxName {
+		nameW = maxName
 	}
-	nameW = avail / 2
-	sparkW = avail - nameW - 2 // -2 for spark column padding
 	if nameW < 13 {
 		nameW = 13
 	}
 	if nameW > 40 {
 		nameW = 40
 	}
+
+	sparkW := avail - nameW
 	if sparkW < 10 {
 		sparkW = 10
 	}
-	if sparkW > 60 {
-		sparkW = 60
+
+	widths[1] = nameW
+	widths[6] = sparkW
+
+	return tableLayout{
+		nameW:     nameW,
+		sparkW:    sparkW,
+		headers:   headers,
+		colWidths: widths,
 	}
-	return
 }

 func (m Model) viewSitesTab() string {
@@ -348,12 +425,16 @@ func (m Model) viewSitesTab() string {
 		return "\n" + welcome
 	}

-	nameW, sparkWidth := m.dynamicWidths()
-	colWidths := []int{6, 0, 10, 10, 8, 8, sparkWidth + 2, 7, 9}
+	layout := m.computeLayout()
+	nameW := layout.nameW
+	sparkWidth := layout.sparkW - 2
+	if sparkWidth < 8 {
+		sparkWidth = 8
+	}

 	var groupRows map[int]bool
 	return m.renderTable(
-		[]string{"#", "NAME", "TYPE", "STATUS", "LATENCY", "UPTIME", "HISTORY", "SSL", "RETRY"},
+		layout.headers,
 		len(m.sites),
 		func(start, end int) [][]string {
 			groupRows = make(map[int]bool)
@@ -366,7 +447,7 @@ func (m Model) viewSitesTab() string {
 					icon := typeIcon("group", m.collapsed[site.ID])
 					rows = append(rows, []string{
 						strconv.Itoa(i + 1),
-						m.zones.Mark(fmt.Sprintf("site-%d", i), icon+" "+limitStr(site.Name, nameW-2)),
+						m.zones.Mark(fmt.Sprintf("site-%d", i), icon+" "+limitStr(site.Name, nameW-4)),
 						"group",
 						fmtStatus(site.Status, site.Paused, m.isMonitorInMaintenance(site.ID)),
 						subtleStyle.Render("—"),
@@ -384,9 +465,17 @@ func (m Model) viewSitesTab() string {
 					if i+1 >= len(m.sites) || m.sites[i+1].ParentID != site.ParentID {
 						prefix = "└"
 					}
-					name = prefix + " " + limitStr(name, nameW-2)
+					name = prefix + " " + limitStr(name, nameW-4)
 				} else {
-					name = limitStr(name, nameW)
+					name = limitStr(name, nameW-2)
+				}
+
+				if (site.Status == "DOWN" || site.Status == "SSL EXP" || site.Status == "LATE") && site.LastError != "" {
+					nameLen := len([]rune(name))
+					errSpace := nameW - nameLen - 3
+					if errSpace > 10 {
+						name = name + " " + subtleStyle.Render(limitStr(site.LastError, errSpace))
+					}
 				}

 				hist, _ := m.engine.GetHistory(site.ID)
@@ -394,7 +483,7 @@ func (m Model) viewSitesTab() string {
 				if site.Type == "push" {
 					spark = heartbeatSparkline(hist.Statuses, sparkWidth)
 				} else {
-					spark = latencySparkline(hist.Latencies, sparkWidth)
+					spark = latencySparkline(hist.Latencies, hist.Statuses, sparkWidth)
 				}

 				rows = append(rows, []string{
@@ -411,7 +500,7 @@ func (m Model) viewSitesTab() string {
 			}
 			return rows
 		},
-		colWidths,
+		layout.colWidths,
 		func(row, col int) *lipgloss.Style {
 			if groupRows[row] {
 				s := siteGroupStyle
@@ -731,7 +820,30 @@ func (m Model) viewDetailPanel() string {
 		fmt.Fprintf(&b, "  %-16s %s\n", subtleStyle.Render(label), value)
 	}

+	section := func(label string) {
+		b.WriteString("\n" + subtleStyle.Render("  "+label) + "\n")
+	}
+
 	row("Status", fmtStatus(site.Status, site.Paused, m.isMonitorInMaintenance(site.ID)))
+
+	if (site.Status == "DOWN" || site.Status == "SSL EXP" || site.Status == "LATE") && site.LastError != "" {
+		row("Error", dangerStyle.Render(limitStr(site.LastError, 60)))
+	}
+
+	if site.Type == "http" && site.StatusCode > 0 {
+		row("HTTP Code", strconv.Itoa(site.StatusCode))
+	}
+
+	if !site.StatusChangedAt.IsZero() {
+		dur := time.Since(site.StatusChangedAt)
+		row("State Since", site.StatusChangedAt.Format("2006-01-02 15:04:05")+" ("+fmtDuration(dur)+")")
+	}
+
+	if !site.LastSuccessAt.IsZero() {
+		ago := time.Since(site.LastSuccessAt)
+		row("Last Success", site.LastSuccessAt.Format("15:04:05")+" ("+fmtDuration(ago)+" ago)")
+	}
+
 	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) {
@@ -740,6 +852,8 @@ func (m Model) viewDetailPanel() string {
 			}
 		}
 	}
+
+	section("ENDPOINT")
 	row("Type", site.Type)
 	if site.URL != "" {
 		row("URL", site.URL)
@@ -750,31 +864,45 @@ func (m Model) viewDetailPanel() string {
 	if site.Port > 0 {
 		row("Port", strconv.Itoa(site.Port))
 	}
+
+	section("TIMING")
 	row("Interval", fmt.Sprintf("%ds", site.Interval))
-	row("Timeout", fmt.Sprintf("%ds", site.Timeout))
+	if site.Timeout > 0 {
+		row("Timeout", fmt.Sprintf("%ds", site.Timeout))
+	}
 	row("Latency", fmtLatency(site.Latency))
 	row("Uptime", fmtUptime(hist.Statuses))
+	if !site.LastCheck.IsZero() {
+		row("Last Check", site.LastCheck.Format("15:04:05"))
+	}

 	if site.Type == "http" {
-		row("Method", site.Method)
-		row("Codes", site.AcceptedCodes)
+		section("HTTP")
+		if site.Method != "" && site.Method != "GET" {
+			row("Method", site.Method)
+		}
+		codes := site.AcceptedCodes
+		if codes == "" {
+			codes = "200-299"
+		}
+		row("Codes", codes)
 		row("SSL", fmtSSL(site))
 		if site.IgnoreTLS {
 			row("TLS Verify", dangerStyle.Render("disabled"))
 		}
 	}

-	if site.MaxRetries > 0 {
-		row("Retries", fmtRetries(site))
-	}
-	if site.Regions != "" {
-		row("Regions", site.Regions)
-	}
-	if site.Description != "" {
-		row("Description", site.Description)
-	}
-	if !site.LastCheck.IsZero() {
-		row("Last Check", site.LastCheck.Format("15:04:05"))
+	if site.MaxRetries > 0 || site.Regions != "" || site.Description != "" {
+		section("CONFIG")
+		if site.MaxRetries > 0 {
+			row("Retries", fmtRetries(site))
+		}
+		if site.Regions != "" {
+			row("Regions", site.Regions)
+		}
+		if site.Description != "" {
+			row("Description", site.Description)
+		}
 	}

 	probeResults := m.engine.GetProbeResults(site.ID)
@@ -787,7 +915,30 @@ func (m Model) viewDetailPanel() string {
 			}
 			latency := time.Duration(result.LatencyNs).Milliseconds()
 			ago := time.Since(result.CheckedAt).Truncate(time.Second)
-			fmt.Fprintf(&b, "  %-14s %s  %dms  %s ago\n", nodeID, status, latency, ago)
+			line := fmt.Sprintf("  %-14s %s  %dms  %s ago", nodeID, status, latency, ago)
+			if !result.IsUp && result.ErrorReason != "" {
+				line += "  " + dangerStyle.Render(limitStr(result.ErrorReason, 30))
+			}
+			b.WriteString(line + "\n")
+		}
+	}
+
+	stateChanges := m.engine.GetStateChanges(site.ID, 5)
+	if len(stateChanges) > 0 {
+		b.WriteString("\n" + subtleStyle.Render("  STATE CHANGES") + "\n")
+		for _, sc := range stateChanges {
+			ago := fmtDuration(time.Since(sc.ChangedAt))
+			arrow := subtleStyle.Render(sc.FromStatus) + " → "
+			if sc.ToStatus == "UP" {
+				arrow += specialStyle.Render(sc.ToStatus)
+			} else {
+				arrow += dangerStyle.Render(sc.ToStatus)
+			}
+			line := fmt.Sprintf("  %s  %s", arrow, subtleStyle.Render(ago+" ago"))
+			if sc.ErrorReason != "" && sc.ToStatus != "UP" {
+				line += "  " + dangerStyle.Render(limitStr(sc.ErrorReason, 40))
+			}
+			b.WriteString(line + "\n")
 		}
 	}

@@ -807,20 +958,27 @@ func (m Model) viewDetailPanel() string {
 				up, len(hist.Statuses))
 		}
 	} else {
-		b.WriteString("  " + latencySparkline(hist.Latencies, sparkWidth))
-		if len(hist.Latencies) > 0 {
-			minL, maxL := hist.Latencies[0], hist.Latencies[0]
-			var total time.Duration
-			for _, l := range hist.Latencies {
-				total += l
-				if l < minL {
-					minL = l
-				}
-				if l > maxL {
-					maxL = l
-				}
+		b.WriteString("  " + latencySparkline(hist.Latencies, hist.Statuses, sparkWidth))
+		// Stats over successful checks only — a failed check is stored as 0ns latency
+		// and would otherwise drag Min to 0ms and skew the average.
+		var minL, maxL, total time.Duration
+		count := 0
+		for i, l := range hist.Latencies {
+			if i < len(hist.Statuses) && !hist.Statuses[i] {
+				continue
 			}
-			avg := total / time.Duration(len(hist.Latencies))
+			if count == 0 {
+				minL, maxL = l, l
+			} else if l < minL {
+				minL = l
+			} else if l > maxL {
+				maxL = l
+			}
+			total += l
+			count++
+		}
+		if count > 0 {
+			avg := total / time.Duration(count)
 			fmt.Fprintf(&b, "\n  %s %dms  %s %dms  %s %dms",
 				subtleStyle.Render("Min"), minL.Milliseconds(),
 				subtleStyle.Render("Avg"), avg.Milliseconds(),
@@ -32,8 +32,19 @@ func (m Model) viewUsersTab() string {
 		return "\n  No users configured. Press [n] to add one."
 	}

+	var headers []string
+	var widths []int
+	if m.isWide() {
+		headers = []string{"#", "USERNAME", "ROLE", "PUBLIC KEY"}
+		widths = []int{4, 18, 10, 50}
+	} else {
+		headers = []string{"#", "USER", "ROLE", "KEY"}
+		widths = []int{4, 14, 8, 30}
+	}
+	userW := widths[1]
+
 	return m.renderTable(
-		[]string{"#", "USERNAME", "ROLE", "PUBLIC KEY"},
+		headers,
 		len(m.users),
 		func(start, end int) [][]string {
 			var rows [][]string
@@ -41,14 +52,14 @@ func (m Model) viewUsersTab() string {
 				u := m.users[i]
 				rows = append(rows, []string{
 					fmt.Sprintf("%d", i+1),
-					m.zones.Mark(fmt.Sprintf("user-%d", i), limitStr(u.Username, 15)),
+					m.zones.Mark(fmt.Sprintf("user-%d", i), limitStr(u.Username, userW-2)),
 					fmtRole(u.Role),
 					fmtKey(u.PublicKey),
 				})
 			}
 			return rows
 		},
-		nil, nil,
+		widths, nil,
 	)
 }

@@ -15,6 +15,12 @@ var (

 type StyleOverride func(row, col int) *lipgloss.Style

+const wideBreakpoint = 120
+
+func (m Model) isWide() bool {
+	return m.termWidth >= wideBreakpoint
+}
+
 func (m Model) renderTable(headers []string, items int, buildRows func(start, end int) [][]string, colWidths []int, styleOverride StyleOverride) string {
 	if items == 0 {
 		return ""
@@ -28,7 +34,16 @@ func (m Model) renderTable(headers []string, items int, buildRows func(start, en
 	selectedVisual := m.cursor - m.tableOffset
 	rows := buildRows(m.tableOffset, end)

-	tableWidth := m.termWidth - chromePadH - 2
+	colTotal := 0
+	for _, w := range colWidths {
+		colTotal += w
+	}
+	borderOverhead := 2 + len(colWidths) - 1
+	tableWidth := colTotal + borderOverhead
+	maxWidth := m.termWidth - chromePadH - 2
+	if tableWidth > maxWidth {
+		tableWidth = maxWidth
+	}
 	if tableWidth < 40 {
 		tableWidth = 40
 	}
@@ -41,7 +56,11 @@ func (m Model) renderTable(headers []string, items int, buildRows func(start, en
 		Rows(rows...).
 		StyleFunc(func(row, col int) lipgloss.Style {
 			if row == table.HeaderRow {
-				return tableHeaderStyle
+				h := tableHeaderStyle
+				if col < len(colWidths) && colWidths[col] > 0 {
+					h = h.Width(colWidths[col]).MaxWidth(colWidths[col])
+				}
+				return h
 			}
 			isSelected := row == selectedVisual
 			if styleOverride != nil {
@@ -51,7 +70,7 @@ func (m Model) renderTable(headers []string, items int, buildRows func(start, en
 						style = tableSelectedStyle.Foreground(s.GetForeground())
 					}
 					if col < len(colWidths) && colWidths[col] > 0 {
-						style = style.Width(colWidths[col])
+						style = style.Width(colWidths[col]).MaxWidth(colWidths[col])
 					}
 					return style
 				}
@@ -64,7 +83,7 @@ func (m Model) renderTable(headers []string, items int, buildRows func(start, en
 				base = tableSelectedStyle
 			}
 			if col < len(colWidths) && colWidths[col] > 0 {
-				base = base.Width(colWidths[col])
+				base = base.Width(colWidths[col]).MaxWidth(colWidths[col])
 			}
 			return base
 		})
@@ -4,6 +4,7 @@ import (
 	"encoding/json"
 	"fmt"
 	"math"
+	"os"
 	"sort"
 	"strings"
 	"time"
@@ -68,6 +69,7 @@ const (
 	stateLogs
 	stateUsers
 	stateDetail
+	stateAlertDetail
 	stateFormSite
 	stateFormAlert
 	stateFormUser
@@ -92,9 +94,10 @@ type Model struct {
 	userFormData  *userFormData
 	maintFormData *maintFormData

-	logViewport viewport.Model
-	isAdmin     bool
-	zones       *zone.Manager
+	logViewport        viewport.Model
+	logFilterImportant bool
+	isAdmin            bool
+	zones              *zone.Manager

 	deleteID   int
 	deleteName string
@@ -120,6 +123,10 @@ type Model struct {

 	filterMode bool
 	filterText string
+
+	// demoMode renders a stable status dot instead of the animated pulse so
+	// screenshots/recordings don't capture the spinner mid-frame. Set via UPTOP_DEMO=1.
+	demoMode bool
 }

 func InitialModel(isAdmin bool, s store.Store, eng *monitor.Engine) Model {
@@ -153,6 +160,7 @@ func InitialModel(isAdmin bool, s store.Store, eng *monitor.Engine) Model {
 		collapsed:    collapsed,
 		theme:        theme,
 		themeIndex:   themeIdx,
+		demoMode:     os.Getenv("UPTOP_DEMO") == "1",
 	}
 }

@@ -383,6 +391,14 @@ func (m Model) Update(msg tea.Msg) (tea.Model, tea.Cmd) {
 				return m, tea.Quit
 			}
 			return m, nil
+		case stateAlertDetail:
+			switch msg.String() {
+			case "i", "esc":
+				m.state = stateDashboard
+			case "q":
+				return m, tea.Quit
+			}
+			return m, nil
 		case stateDashboard, stateLogs, stateUsers:
 			switch msg.String() {
 			case "q":
@@ -392,6 +408,11 @@ func (m Model) Update(msg tea.Msg) (tea.Model, tea.Cmd) {
 					m.filterMode = true
 					return m, nil
 				}
+			case "f":
+				if m.state == stateLogs {
+					m.logFilterImportant = !m.logFilterImportant
+					return m, nil
+				}
 			case "tab":
 				m.switchTab(m.currentTab + 1)
 			case "pgup", "pgdown":
@@ -463,6 +484,16 @@ func (m Model) Update(msg tea.Msg) (tea.Model, tea.Cmd) {
 					m.state = stateFormUser
 					return m, m.initUserHuhForm()
 				}
+			case "t":
+				if m.currentTab == 1 && len(m.alerts) > 0 {
+					a := m.alerts[m.cursor]
+					go func() {
+						if err := m.engine.TestAlert(a.ID); err != nil {
+							m.engine.AddLog(fmt.Sprintf("Test alert failed (%s): %v", a.Name, err))
+						}
+					}()
+					return m, nil
+				}
 			case " ":
 				if m.currentTab == 0 && len(m.sites) > 0 && m.sites[m.cursor].Type == "group" {
 					gid := m.sites[m.cursor].ID
@@ -481,6 +512,8 @@ func (m Model) Update(msg tea.Msg) (tea.Model, tea.Cmd) {
 			case "i":
 				if m.currentTab == 0 && len(m.sites) > 0 {
 					m.state = stateDetail
+				} else if m.currentTab == 1 && len(m.alerts) > 0 {
+					m.state = stateAlertDetail
 				}
 			case "x":
 				if m.currentTab == 4 && len(m.maintenanceWindows) > 0 {
@@ -727,11 +760,6 @@ func (m *Model) submitForm() {
 }

 func (m Model) pulseIndicator() string {
-	frame := m.tickCount % len(pulseFrames)
-	brightness := int(m.pulsePos*155) + 100
-	if brightness > 255 {
-		brightness = 255
-	}
 	hasDown := false
 	for _, s := range m.sites {
 		if !s.Paused && !m.isMonitorInMaintenance(s.ID) && (s.Status == "DOWN" || s.Status == "SSL EXP") {
@@ -739,6 +767,19 @@ func (m Model) pulseIndicator() string {
 			break
 		}
 	}
+	// Stills can't show animation: render a stable status dot in demo mode.
+	if m.demoMode {
+		c := m.theme.Success
+		if hasDown {
+			c = m.theme.Danger
+		}
+		return lipgloss.NewStyle().Foreground(c).Render("●")
+	}
+	frame := m.tickCount % len(pulseFrames)
+	brightness := int(m.pulsePos*155) + 100
+	if brightness > 255 {
+		brightness = 255
+	}
 	var color string
 	if hasDown {
 		color = fmt.Sprintf("#%02x%02x%02x", brightness, brightness/4, brightness/4)
@@ -802,6 +843,8 @@ func (m Model) View() string {
 		return ""
 	case stateDetail:
 		return m.viewDetailPanel()
+	case stateAlertDetail:
+		return m.viewAlertDetailPanel()
 	default:
 		return m.zones.Scan(m.viewDashboard())
 	}
@@ -811,13 +854,20 @@ func (m Model) viewDashboard() string {
 	allSites := m.engine.GetAllSites()
 	totalMonitors := 0
 	downCount := 0
+	lateCount := 0
 	for _, s := range allSites {
 		if s.Type == "group" {
 			continue
 		}
 		totalMonitors++
-		if !s.Paused && !m.isMonitorInMaintenance(s.ID) && (s.Status == "DOWN" || s.Status == "SSL EXP") {
+		if s.Paused || m.isMonitorInMaintenance(s.ID) {
+			continue
+		}
+		switch s.Status {
+		case "DOWN", "SSL EXP":
 			downCount++
+		case "LATE":
+			lateCount++
 		}
 	}
 	offlineNodes := 0
@@ -830,6 +880,8 @@ func (m Model) viewDashboard() string {
 	var sitesLabel string
 	if downCount > 0 {
 		sitesLabel = fmt.Sprintf("Sites (%d↓)", downCount)
+	} else if lateCount > 0 {
+		sitesLabel = fmt.Sprintf("Sites (%d⚠)", lateCount)
 	} else if totalMonitors > 0 {
 		sitesLabel = fmt.Sprintf("Sites (%d)", totalMonitors)
 	} else {
@@ -895,14 +947,19 @@ func (m Model) viewDashboard() string {
 		}
 	}

-	upCount := totalMonitors - downCount
+	upCount := totalMonitors - downCount - lateCount
 	var upStr string
 	if downCount > 0 {
 		upStr = dangerStyle.Render(fmt.Sprintf("%d/%d UP", upCount, totalMonitors))
+	} else if lateCount > 0 {
+		upStr = warnStyle.Render(fmt.Sprintf("%d/%d UP", upCount, totalMonitors))
 	} else {
 		upStr = specialStyle.Render(fmt.Sprintf("%d/%d UP", upCount, totalMonitors))
 	}
 	statusParts := []string{upStr}
+	if lateCount > 0 {
+		statusParts = append(statusParts, warnStyle.Render(fmt.Sprintf("%d LATE", lateCount)))
+	}
 	if len(m.nodes) > 0 {
 		online := 0
 		for _, n := range m.nodes {
@@ -910,7 +967,11 @@ func (m Model) viewDashboard() string {
 				online++
 			}
 		}
-		statusParts = append(statusParts, fmt.Sprintf("%d probes", online))
+		probeLabel := "probes"
+		if online == 1 {
+			probeLabel = "probe"
+		}
+		statusParts = append(statusParts, fmt.Sprintf("%d %s", online, probeLabel))
 	}
 	statusLine := strings.Join(statusParts, subtleStyle.Render(" · "))

@@ -923,6 +984,10 @@ func (m Model) viewDashboard() string {
 		switch m.currentTab {
 		case 0:
 			keys = "[/]Filter [n]New [e]Edit [i]Info [d]Del [p]Pause [T]Theme [Tab]Switch [q]Quit"
+		case 1:
+			keys = "[n]New [e]Edit [i]Info [d]Del [t]Test [T]Theme [Tab]Switch [q]Quit"
+		case 2:
+			keys = "[f]Filter [T]Theme [Tab]Switch [q]Quit"
 		case 4:
 			keys = "[n]New [x]End [d]Del [T]Theme [Tab]Switch [q]Quit"
 		case 5:
@@ -949,10 +1014,12 @@ func siteOrder(s models.Site) int {
 	switch s.Status {
 	case "DOWN", "SSL EXP":
 		return 0
-	case "PENDING":
-		return 2
-	default:
+	case "LATE":
 		return 1
+	case "PENDING":
+		return 3
+	default:
+		return 2
 	}
 }
Author	SHA1	Message	Date
lerko	4c79f93d53	chore: add TUI screenshots CI / test (pull_request) Successful in 2m50s Details CI / lint (pull_request) Successful in 1m7s Details CI / vulncheck (pull_request) Successful in 56s Details Six tab screenshots: monitors, alerts, logs, nodes, detail, theme. Generated with VHS; tooling lives in lerko/uptop-vhs.	2026-05-29 12:35:15 -04:00
lerko	9ce6153bd8	feat(tui): add detail panel, type icons, and store improvements Wire up monitor detail view in TUI with type-specific icons. Add SQLite WAL glob to gitignore. Extend store interface with bulk-fetch and history queries for the detail panel.	2026-05-29 12:35:08 -04:00
lerko	cfbf01274d	chore(tui): visual polish — detail sections, column headers, alert detail (#37 ) CI / test (push) Successful in 2m40s Details CI / lint (push) Successful in 1m2s Details CI / vulncheck (push) Successful in 51s Details Release / docker (push) Has been cancelled Details Release / release (push) Has been cancelled Details ## Summary Bundled remaining UX polish items from the screenshot review. ### Changes Detail panel sections (#5) - Fields grouped into ENDPOINT, TIMING, HTTP, CONFIG sections with subtle headers - Matches existing PROBE RESULTS and STATE CHANGES section pattern - Cleaner visual hierarchy without box-drawing clutter Omit unconfigured fields (#6) - Timeout hidden when 0 (unconfigured) - Method hidden when default GET - AcceptedCodes shows "200-299" explicitly when empty Column header (#7) - `LATENCY` → `LAT` (design short, never truncate — htop/btop pattern) Alert detail view (#8) - `i` key on Alerts tab opens full detail panel - Shows: type, health status, last sent time, send/fail counts, last error - Full config key:value pairs (untruncated) - Keybinding: `[i/Esc] Back [e] Edit [t] Test [q] Quit` ### Files (3) - `internal/tui/tab_sites.go` — section headers, field omission, LAT header - `internal/tui/tab_alerts.go` — viewAlertDetailPanel() - `internal/tui/tui.go` — stateAlertDetail, key handler, render routing Reviewed-on: lerko/uptop#37	2026-05-28 20:40:29 +00:00
lerko	26e297cbae	Merge pull request 'feat: alert channel health indicator + test alerts' (#36 ) from feat/alert-health into main CI / test (push) Successful in 2m48s Details CI / lint (push) Successful in 1m17s Details CI / vulncheck (push) Successful in 1m6s Details Reviewed-on: lerko/uptop#36	2026-05-28 01:33:00 +00:00
lerko	0aa2f9cd8a	feat: alert channel health indicator + test alerts CI / test (pull_request) Successful in 2m46s Details CI / lint (pull_request) Successful in 1m1s Details CI / vulncheck (pull_request) Successful in 51s Details Track alert delivery health at runtime: - AlertHealth struct: LastSendAt, LastSendOK, LastError, SendCount, FailCount - triggerAlert records success/failure after each Send() - Health data exposed via GetAlertHealth() for TUI Alerts tab enriched: - Health dot column: green (OK), red (failed), gray (never sent) - LAST SENT column: relative time ("2m ago", "never") - [t] key sends test notification through selected channel Inspired by Grafana's contact point health columns.	2026-05-27 21:23:06 -04:00
lerko	f17f06a1c6	Merge pull request 'feat: logs tab overhaul — severity tags, filtering, recovery durations' (#35 ) from feat/logs-overhaul into main CI / test (push) Successful in 2m47s Details CI / lint (push) Successful in 1m16s Details CI / vulncheck (push) Successful in 56s Details Reviewed-on: lerko/uptop#35	2026-05-28 00:35:24 +00:00
lerko	b14d5e19db	feat: logs tab overhaul — severity tags, filtering, recovery durations CI / test (pull_request) Successful in 2m36s Details CI / lint (pull_request) Successful in 1m1s Details CI / vulncheck (pull_request) Successful in 51s Details Logs tab visual overhaul: - Severity-classified entries: DOWN (red), UP (green), WARN (amber), SYS (cyan), info (gray) — rendered as inline tags, not whole-line color - Column-aligned format: [timestamp] [severity tag] [message] - Filter toggle (f key): All vs Important only (hides retry noise) - Header shows entry count, filter state, hidden count Engine log improvements: - Recovery messages include downtime duration ("was down 14m") - LATE transition logged ("heartbeat overdue") - Push monitor recovery includes downtime duration	2026-05-27 20:14:43 -04:00
lerko	a2b38ddc60	Merge pull request 'feat: proper push monitor lifecycle — PENDING, LATE, DOWN' (#34 ) from feat/push-monitor-states into main CI / test (push) Successful in 2m48s Details CI / lint (push) Successful in 1m17s Details CI / vulncheck (push) Successful in 56s Details Reviewed-on: lerko/uptop#34	2026-05-28 00:01:56 +00:00
lerko	5dc31108f8	feat: proper push monitor lifecycle — PENDING, LATE, DOWN states CI / test (pull_request) Successful in 2m41s Details CI / lint (pull_request) Successful in 1m7s Details CI / vulncheck (pull_request) Successful in 46s Details Push monitors no longer lie about status: - PENDING stays until first heartbeat (no auto-promote to UP) - LATE state (amber) when overdue but within grace period - DOWN only after grace period expires - Grace period = interval/2, minimum 60s RecordHeartbeat now handles all transitions: - PENDING → UP (first heartbeat, logged) - LATE → UP (late arrival, logged) - DOWN → UP (recovery, alert + state change persisted) TUI updates: - LATE rendered in amber/warning color - Status bar shows LATE count separately - Tab badge shows ⚠ for late monitors - Sort order: DOWN > LATE > UP > PENDING > PAUSED - Detail panel shows error for LATE monitors Inspired by Healthchecks.io state machine (new/up/grace/down).	2026-05-27 19:56:50 -04:00
lerko	63773b13d0	Merge pull request 'feat: show error reason when monitors go DOWN' (#33 ) from feat/error-reason into main CI / test (push) Successful in 2m51s Details CI / lint (push) Successful in 1m1s Details CI / vulncheck (push) Successful in 56s Details Reviewed-on: lerko/uptop#33	2026-05-27 23:38:26 +00:00
lerko	bc3a44beac	feat: show error reason when monitors go DOWN CI / test (pull_request) Successful in 2m42s Details CI / lint (pull_request) Successful in 1m11s Details CI / vulncheck (pull_request) Successful in 51s Details Propagate check failure reasons through the entire stack: - Checker captures specific errors (DNS, timeout, HTTP status, SSL, etc.) - Engine tracks LastError, StatusChangedAt, LastSuccessAt per monitor - State transitions persisted to new state_changes table - Detail panel shows error reason, HTTP code, state duration, last success time, and last 5 state change events - Monitor table shows inline error preview for DOWN services - Alert messages include error reason - Probe nodes forward error reasons to leader 15 files changed across models, checker, engine, store, TUI, and probes.	2026-05-27 19:32:30 -04:00