feat: infra backup GFS retention + version history

New infra_backup_versions table with GFS pruning (~14 versions per
customer). Recovery endpoint supports ?version=ID. New /versions API.
Dashboard shows collapsible backup history with app names and disk count.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

hub/CHANGELOG.md

@@ -1,5 +1,17 @@
 # Felhom Hub — Changelog
 
+## v0.6.2 (2026-02-26)
+
+### Added
+- **Infra backup GFS retention** — New `infra_backup_versions` table stores multiple backups per customer. GFS pruning keeps: all from last 24h, latest per day (7 days), latest per week (4 weeks), latest per month (3 months) — ~14 versions max per customer
+- **`GET /api/v1/infra-backup/{id}/versions`** — Returns metadata list of all retained backup versions (date, stack names, disk count) for a customer. Bearer auth.
+- **Recovery version selection** — `GET /api/v1/recovery/{id}?version=ID` fetches a specific backup version instead of latest. Response now includes `backup_versions` array with all available versions.
+- **Dashboard backup history** — Customer detail page "Infra Backup" card shows version count and collapsible history table (date, apps, disks)
+
+### Changed
+- **`SaveInfraBackup()`** — Now INSERTs a new row instead of upserting, preserving history. Automatically prunes old versions via GFS algorithm.
+- **One-time migration** — Existing data from `infra_backups` table is copied to `infra_backup_versions` on first startup
+
 ## v0.6.1 (2026-02-25)
 
 ### Added

hub/internal/api/handler.go

@@ -102,6 +102,10 @@ func (h *Handler) ServeHTTP(w http.ResponseWriter, r *http.Request) {
 		h.handleNotify(w, r)
 	case r.Method == http.MethodPost && path == "/infra-backup":
 		h.handleInfraBackupPush(w, r)
+	case r.Method == http.MethodGet && strings.HasPrefix(path, "/infra-backup/") && strings.HasSuffix(path, "/versions"):
+		customerID := strings.TrimPrefix(path, "/infra-backup/")
+		customerID = strings.TrimSuffix(customerID, "/versions")
+		h.handleInfraBackupVersions(w, r, customerID)
 	case r.Method == http.MethodGet && strings.HasPrefix(path, "/infra-backup/"):
 		h.handleInfraBackupGet(w, r, strings.TrimPrefix(path, "/infra-backup/"))
 	case r.Method == http.MethodPost && path == "/preferences":
@@ -586,6 +590,33 @@ func (h *Handler) handleInfraBackupGet(w http.ResponseWriter, r *http.Request, c
 	w.Write(data)
 }
 
+// handleInfraBackupVersions returns a list of backup versions for a customer.
+// Auth: Bearer token.
+func (h *Handler) handleInfraBackupVersions(w http.ResponseWriter, r *http.Request, customerID string) {
+	if !h.checkAuth(r) {
+		http.Error(w, "Unauthorized", http.StatusUnauthorized)
+		return
+	}
+
+	if customerID == "" {
+		http.Error(w, "Missing customer_id", http.StatusBadRequest)
+		return
+	}
+
+	versions, err := h.store.ListInfraBackupVersions(customerID)
+	if err != nil {
+		h.logger.Printf("[ERROR] Failed to list infra backup versions for %s: %v", customerID, err)
+		http.Error(w, "Internal error", http.StatusInternalServerError)
+		return
+	}
+	if versions == nil {
+		versions = []store.InfraBackupVersion{}
+	}
+
+	w.Header().Set("Content-Type", "application/json")
+	json.NewEncoder(w).Encode(versions)
+}
+
 // handleRecovery returns both the generated controller.yaml and the infra backup for disaster recovery.
 // Auth: X-Retrieval-Password header (same as config retrieval).
 func (h *Handler) handleRecovery(w http.ResponseWriter, r *http.Request, customerID string) {
@@ -631,24 +662,44 @@ func (h *Handler) handleRecovery(w http.ResponseWriter, r *http.Request, custome
 	// Fetch infra backup (optional — may not exist for new customers)
 	var infraBackup json.RawMessage
 	hasInfraBackup := false
+
+	// Support ?version=ID for selecting a specific backup version
+	if versionStr := r.URL.Query().Get("version"); versionStr != "" {
+		var versionID int64
+		if _, err := fmt.Sscanf(versionStr, "%d", &versionID); err == nil {
+			if data, err := h.store.GetInfraBackupByID(versionID); err == nil && data != nil {
+				infraBackup = data
+				hasInfraBackup = true
+			}
+		}
+	} else {
 		if data, err := h.store.GetInfraBackup(customerID); err == nil && data != nil {
 			infraBackup = data
 			hasInfraBackup = true
 		}
+	}
+
+	// Include version list for version picker
+	var backupVersions []store.InfraBackupVersion
+	if versions, err := h.store.ListInfraBackupVersions(customerID); err == nil {
+		backupVersions = versions
+	}
 
 	resp := struct {
 		CustomerID     string                    `json:"customer_id"`
 		ConfigYAML     string                    `json:"config_yaml"`
 		InfraBackup    json.RawMessage           `json:"infra_backup"`
 		HasInfraBackup bool                      `json:"has_infra_backup"`
+		BackupVersions []store.InfraBackupVersion `json:"backup_versions,omitempty"`
 	}{
 		CustomerID:     customerID,
 		ConfigYAML:     configYAML,
 		InfraBackup:    infraBackup,
 		HasInfraBackup: hasInfraBackup,
+		BackupVersions: backupVersions,
 	}
 
-	h.logger.Printf("[INFO] Recovery data downloaded for customer %s (has_infra_backup=%v)", customerID, hasInfraBackup)
+	h.logger.Printf("[INFO] Recovery data downloaded for customer %s (has_infra_backup=%v, versions=%d)", customerID, hasInfraBackup, len(backupVersions))
 	w.Header().Set("Content-Type", "application/json")
 	json.NewEncoder(w).Encode(resp)
 }

hub/internal/store/store.go

@@ -195,6 +195,27 @@ func (s *Store) migrate() error {
 		return err
 	}
 
+	// v0.7.0: versioned infra backups with GFS retention
+	_, err = s.db.Exec(`
+		CREATE TABLE IF NOT EXISTS infra_backup_versions (
+			id INTEGER PRIMARY KEY AUTOINCREMENT,
+			customer_id TEXT NOT NULL,
+			backup_json TEXT NOT NULL,
+			created_at DATETIME NOT NULL DEFAULT (datetime('now'))
+		);
+		CREATE INDEX IF NOT EXISTS idx_ibv_customer_time
+			ON infra_backup_versions(customer_id, created_at DESC);
+	`)
+	if err != nil {
+		return err
+	}
+
+	// One-time migration: copy existing single-row backups to versioned table
+	s.db.Exec(`INSERT INTO infra_backup_versions (customer_id, backup_json, created_at)
+		SELECT customer_id, backup_json, updated_at FROM infra_backups
+		WHERE NOT EXISTS (SELECT 1 FROM infra_backup_versions
+			WHERE infra_backup_versions.customer_id = infra_backups.customer_id)`)
+
 	return nil
 }
 
@@ -517,24 +538,34 @@ func (s *Store) GetCustomerHistory(customerID string, since time.Duration) ([]Cu
 	return history, rows.Err()
 }
 
-// SaveInfraBackup upserts the infrastructure backup for a customer.
+// SaveInfraBackup inserts a new infra backup version and prunes old ones (GFS retention).
 func (s *Store) SaveInfraBackup(customerID string, backupJSON []byte) error {
 	_, err := s.db.Exec(`
-		INSERT INTO infra_backups (customer_id, backup_json, updated_at)
+		INSERT INTO infra_backup_versions (customer_id, backup_json, created_at)
+		VALUES (?, ?, datetime('now'))`,
+		customerID, string(backupJSON),
+	)
+	if err != nil {
+		return err
+	}
+
+	// Also maintain legacy table for backward compatibility during rollback window
+	s.db.Exec(`INSERT INTO infra_backups (customer_id, backup_json, updated_at)
 		VALUES (?, ?, datetime('now'))
 		ON CONFLICT(customer_id) DO UPDATE SET
 			backup_json = excluded.backup_json,
 			updated_at = datetime('now')`,
-		customerID, string(backupJSON),
+		customerID, string(backupJSON))
-	)
+
-	return err
+	s.pruneInfraBackups(customerID)
+	return nil
 }
 
-// GetInfraBackup returns the raw infra backup JSON for a customer, or nil if not found.
+// GetInfraBackup returns the latest infra backup JSON for a customer, or nil if not found.
 func (s *Store) GetInfraBackup(customerID string) ([]byte, error) {
 	var data string
 	err := s.db.QueryRow(
-		"SELECT backup_json FROM infra_backups WHERE customer_id = ?",
+		"SELECT backup_json FROM infra_backup_versions WHERE customer_id = ? ORDER BY created_at DESC LIMIT 1",
 		customerID,
 	).Scan(&data)
 	if err == sql.ErrNoRows {
@@ -546,20 +577,36 @@ func (s *Store) GetInfraBackup(customerID string) ([]byte, error) {
 	return []byte(data), nil
 }
 
+// GetInfraBackupByID returns the infra backup JSON for a specific version ID, or nil if not found.
+func (s *Store) GetInfraBackupByID(id int64) ([]byte, error) {
+	var data string
+	err := s.db.QueryRow(
+		"SELECT backup_json FROM infra_backup_versions WHERE id = ?", id,
+	).Scan(&data)
+	if err == sql.ErrNoRows {
+		return nil, nil
+	}
+	if err != nil {
+		return nil, err
+	}
+	return []byte(data), nil
+}
+
 // InfraBackupMeta holds summary info for the dashboard (avoids parsing full JSON).
 type InfraBackupMeta struct {
 	UpdatedAt    time.Time
 	StackCount   int
 	DiskCount    int
+	VersionCount int
 }
 
-// GetInfraBackupMeta returns summary metadata for a customer's infra backup.
+// GetInfraBackupMeta returns summary metadata for a customer's latest infra backup.
 func (s *Store) GetInfraBackupMeta(customerID string) (*InfraBackupMeta, error) {
-	var backupJSON, updatedAt string
+	var backupJSON, createdAt string
 	err := s.db.QueryRow(
-		"SELECT backup_json, updated_at FROM infra_backups WHERE customer_id = ?",
+		"SELECT backup_json, created_at FROM infra_backup_versions WHERE customer_id = ? ORDER BY created_at DESC LIMIT 1",
 		customerID,
-	).Scan(&backupJSON, &updatedAt)
+	).Scan(&backupJSON, &createdAt)
 	if err == sql.ErrNoRows {
 		return nil, nil
 	}
@@ -568,24 +615,194 @@ func (s *Store) GetInfraBackupMeta(customerID string) (*InfraBackupMeta, error)
 	}
 
 	meta := &InfraBackupMeta{
-		UpdatedAt: parseSQLiteTime(updatedAt),
+		UpdatedAt: parseSQLiteTime(createdAt),
 	}
 
+	// Count total versions
+	s.db.QueryRow("SELECT COUNT(*) FROM infra_backup_versions WHERE customer_id = ?", customerID).Scan(&meta.VersionCount)
+
 	// Parse just the fields we need
+	parseInfraBackupCounts(backupJSON, &meta.StackCount, &meta.DiskCount, nil, s.logger, customerID)
+
+	return meta, nil
+}
+
+// InfraBackupVersion holds summary info for a single backup version.
+type InfraBackupVersion struct {
+	ID         int64     `json:"id"`
+	CreatedAt  time.Time `json:"created_at"`
+	StackCount int       `json:"stack_count"`
+	DiskCount  int       `json:"disk_count"`
+	StackNames []string  `json:"stack_names,omitempty"`
+}
+
+// ListInfraBackupVersions returns metadata for all retained versions of a customer's backup.
+func (s *Store) ListInfraBackupVersions(customerID string) ([]InfraBackupVersion, error) {
+	rows, err := s.db.Query(
+		"SELECT id, backup_json, created_at FROM infra_backup_versions WHERE customer_id = ? ORDER BY created_at DESC",
+		customerID,
+	)
+	if err != nil {
+		return nil, err
+	}
+	defer rows.Close()
+
+	var versions []InfraBackupVersion
+	for rows.Next() {
+		var v InfraBackupVersion
+		var backupJSON, createdAt string
+		if err := rows.Scan(&v.ID, &backupJSON, &createdAt); err != nil {
+			return nil, err
+		}
+		v.CreatedAt = parseSQLiteTime(createdAt)
+		parseInfraBackupCounts(backupJSON, &v.StackCount, &v.DiskCount, &v.StackNames, nil, "")
+		versions = append(versions, v)
+	}
+	return versions, rows.Err()
+}
+
+// pruneInfraBackups applies GFS retention: keep all from last 24h, latest per day (7d),
+// latest per week (4w), latest per month (3mo). Delete everything else.
+func (s *Store) pruneInfraBackups(customerID string) {
+	rows, err := s.db.Query(
+		"SELECT id, created_at FROM infra_backup_versions WHERE customer_id = ? ORDER BY created_at DESC",
+		customerID,
+	)
+	if err != nil {
+		return
+	}
+	defer rows.Close()
+
+	type entry struct {
+		id        int64
+		createdAt time.Time
+	}
+	var all []entry
+	for rows.Next() {
+		var e entry
+		var ts string
+		if err := rows.Scan(&e.id, &ts); err != nil {
+			return
+		}
+		e.createdAt = parseSQLiteTime(ts)
+		all = append(all, e)
+	}
+
+	if len(all) <= 1 {
+		return
+	}
+
+	now := time.Now().UTC()
+	keep := make(map[int64]bool)
+	seenDays := make(map[string]bool)
+	seenWeeks := make(map[string]bool)
+	seenMonths := make(map[string]bool)
+
+	for _, e := range all {
+		age := now.Sub(e.createdAt)
+
+		// Keep all from last 24h
+		if age < 24*time.Hour {
+			keep[e.id] = true
+			continue
+		}
+
+		// Latest per calendar day for last 7 days
+		if age < 7*24*time.Hour {
+			day := e.createdAt.Format("2006-01-02")
+			if !seenDays[day] {
+				seenDays[day] = true
+				keep[e.id] = true
+			}
+			continue
+		}
+
+		// Latest per ISO week for last 4 weeks
+		if age < 28*24*time.Hour {
+			year, week := e.createdAt.ISOWeek()
+			wk := fmt.Sprintf("%d-W%02d", year, week)
+			if !seenWeeks[wk] {
+				seenWeeks[wk] = true
+				keep[e.id] = true
+			}
+			continue
+		}
+
+		// Latest per calendar month for last 3 months
+		if age < 90*24*time.Hour {
+			month := e.createdAt.Format("2006-01")
+			if !seenMonths[month] {
+				seenMonths[month] = true
+				keep[e.id] = true
+			}
+			continue
+		}
+
+		// Older than 3 months — don't keep
+	}
+
+	// Build delete list
+	var deleteIDs []interface{}
+	var placeholders []string
+	for _, e := range all {
+		if !keep[e.id] {
+			deleteIDs = append(deleteIDs, e.id)
+			placeholders = append(placeholders, "?")
+		}
+	}
+
+	if len(deleteIDs) == 0 {
+		return
+	}
+
+	query := "DELETE FROM infra_backup_versions WHERE id IN (" + joinStrings(placeholders, ",") + ")"
+	_, err = s.db.Exec(query, deleteIDs...)
+	if err != nil {
+		s.logger.Printf("[WARN] Failed to prune infra backup versions for %s: %v", customerID, err)
+	} else {
+		s.logger.Printf("[INFO] Pruned %d old infra backup version(s) for %s (kept %d)", len(deleteIDs), customerID, len(keep))
+	}
+}
+
+// parseInfraBackupCounts extracts stack/disk counts and optionally stack names from backup JSON.
+func parseInfraBackupCounts(backupJSON string, stackCount, diskCount *int, stackNames *[]string, logger *log.Logger, customerID string) {
 	var parsed struct {
-		DeployedStacks []json.RawMessage `json:"deployed_stacks"`
+		DeployedStacks []struct {
+			Name        string `json:"name"`
+			DisplayName string `json:"display_name"`
+		} `json:"deployed_stacks"`
 		DiskLayout struct {
 			Mounts []json.RawMessage `json:"mounts"`
 		} `json:"disk_layout"`
 	}
 	if err := json.Unmarshal([]byte(backupJSON), &parsed); err != nil {
-		s.logger.Printf("[WARN] Failed to parse infra backup metadata for %s: %v", customerID, err)
+		if logger != nil {
-	} else {
+			logger.Printf("[WARN] Failed to parse infra backup metadata for %s: %v", customerID, err)
-		meta.StackCount = len(parsed.DeployedStacks)
-		meta.DiskCount = len(parsed.DiskLayout.Mounts)
 		}
+		return
+	}
+	*stackCount = len(parsed.DeployedStacks)
+	*diskCount = len(parsed.DiskLayout.Mounts)
+	if stackNames != nil {
+		for _, s := range parsed.DeployedStacks {
+			name := s.DisplayName
+			if name == "" {
+				name = s.Name
+			}
+			*stackNames = append(*stackNames, name)
+		}
+	}
+}
 
-	return meta, nil
+func joinStrings(ss []string, sep string) string {
+	if len(ss) == 0 {
+		return ""
+	}
+	result := ss[0]
+	for _, s := range ss[1:] {
+		result += sep + s
+	}
+	return result
 }
 
 // Prune deletes reports older than the given number of days.

hub/internal/web/configs.go

@@ -231,6 +231,7 @@ func (s *Server) handleCustomerUnified(w http.ResponseWriter, r *http.Request, c
 	var recentNotifs []store.NotificationLogEntry
 	var infraMeta *store.InfraBackupMeta
 	var infraBackupAge string
+	var infraBackupVersions []store.InfraBackupVersion
 	var events []store.Event
 	var eventCounts map[string]int
 
@@ -244,6 +245,7 @@ func (s *Server) handleCustomerUnified(w http.ResponseWriter, r *http.Request, c
 		if infraMeta != nil {
 			infraBackupAge = timeAgo(infraMeta.UpdatedAt)
 		}
+		infraBackupVersions, _ = s.store.ListInfraBackupVersions(customerID)
 		events, _ = s.store.GetRecentEvents(customerID, 50)
 		eventCounts, _ = s.store.CountEventsBySeverity(customerID, time.Now().Add(-24*time.Hour))
 		appTelemetry, _ = s.store.GetCustomerAppSummary(customerID, time.Now().Add(-7*24*time.Hour))
@@ -274,6 +276,7 @@ func (s *Server) handleCustomerUnified(w http.ResponseWriter, r *http.Request, c
 
 		InfraBackup         *store.InfraBackupMeta
 		InfraBackupAge      string
+		InfraBackupVersions []store.InfraBackupVersion
 		NotifPrefs          *store.NotificationPrefs
 		RecentNotifications []store.NotificationLogEntry
 		History             []store.CustomerSummary
@@ -315,6 +318,7 @@ func (s *Server) handleCustomerUnified(w http.ResponseWriter, r *http.Request, c
 
 		InfraBackup:         infraMeta,
 		InfraBackupAge:      infraBackupAge,
+		InfraBackupVersions: infraBackupVersions,
 		NotifPrefs:          notifPrefs,
 		RecentNotifications: recentNotifs,
 		History:             history,

hub/internal/web/templates/customer_unified.html

@@ -313,7 +313,34 @@
                     <span class="label">Disks</span>
                     <span class="value">{{.InfraBackup.DiskCount}}</span>
                 </div>
+                <div class="info-item">
+                    <span class="label">Versions</span>
+                    <span class="value">{{.InfraBackup.VersionCount}}</span>
                 </div>
+            </div>
+            {{if .InfraBackupVersions}}
+            <details style="margin-top: 0.75rem;">
+                <summary style="cursor: pointer; color: var(--text-secondary, #94a3b8); font-size: 0.85em;">Backup History ({{len .InfraBackupVersions}} versions)</summary>
+                <table style="width: 100%; margin-top: 0.5rem; font-size: 0.85em;">
+                    <thead>
+                        <tr>
+                            <th style="text-align: left; padding: 0.25rem 0.5rem;">Date</th>
+                            <th style="text-align: left; padding: 0.25rem 0.5rem;">Apps</th>
+                            <th style="text-align: right; padding: 0.25rem 0.5rem;">Disks</th>
+                        </tr>
+                    </thead>
+                    <tbody>
+                    {{range .InfraBackupVersions}}
+                        <tr>
+                            <td style="padding: 0.25rem 0.5rem;">{{.CreatedAt.Format "2006-01-02 15:04"}}</td>
+                            <td style="padding: 0.25rem 0.5rem;">{{.StackCount}}{{if .StackNames}}: {{range $i, $n := .StackNames}}{{if $i}}, {{end}}{{$n}}{{end}}{{end}}</td>
+                            <td style="text-align: right; padding: 0.25rem 0.5rem;">{{.DiskCount}}</td>
+                        </tr>
+                    {{end}}
+                    </tbody>
+                </table>
+            </details>
+            {{end}}
             {{else}}
             <p style="color: #facc15">No infra backup received yet</p>
             {{end}}