feat: logs tab overhaul — severity tags, filtering, recovery durations

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
parent a2b38ddc60
commit b14d5e19db
3 changed files with 134 additions and 30 deletions
@@ -96,6 +96,19 @@ 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()
@@ -206,8 +219,12 @@ func (e *Engine) RecordHeartbeat(token string) bool {
 	case "LATE":
 		e.AddLog(fmt.Sprintf("Push Monitor '%s' heartbeat arrived (was late)", site.Name))
 	case "DOWN":
-		e.AddLog(fmt.Sprintf("Push Monitor '%s' recovered", site.Name))
-		go e.triggerAlert(site.AlertID, "✅ RECOVERY", fmt.Sprintf("Push Monitor '%s' is receiving heartbeats.", site.Name))
+		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" {
@@ -514,6 +531,11 @@ func (e *Engine) handleStatusChange(site models.Site, rawStatus string, code int
 	}

 	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))
@@ -529,11 +551,17 @@ 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))
 	}
 }

@@ -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))
+		total++
+		sev := classifyLog(line)
+		if m.logFilterImportant && !isImportantLog(sev) {
+			continue
+		}
+		shown++
+		rendered = append(rendered, renderLogLine(line))
 	}

-	count := 0
-	for _, l := range lines {
-		if strings.TrimSpace(l) != "" {
-			count++
-		}
+	filterLabel := "All"
+	if m.logFilterImportant {
+		filterLabel = "Important"
 	}

-	header := subtleStyle.Render(fmt.Sprintf("  %d entries  [↑/↓] Scroll  [PgUp/PgDn] Page", count))
-	return "\n" + header + "\n\n" + strings.Join(colored, "\n")
+	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")
 }
@@ -93,6 +93,7 @@ type Model struct {
 	maintFormData *maintFormData

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

@@ -392,6 +393,11 @@ 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":
@@ -937,6 +943,8 @@ 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 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: