Agent-Status und Inaktivitaets-Erkennung implementiert
- Status wird aus last_heartbeat berechnet (online/stale/offline/unknown)
- GET /agents und GET /agents/{id} liefern status Feld
- agent_events Tabelle fuer Event-Tracking (Migration automatisch)
- Inaktivitaets-Monitor als Goroutine (prueft alle 60s, schreibt Events bei Statuswechsel)
- WebSocket Connect/Disconnect Events werden automatisch geschrieben
- API-Endpoints: GET /agents/events und GET /agents/{id}/events (?limit=N&type=X)
- DB-Methoden: InsertAgentEvent, GetAgentEvents, GetAllAgentEvents, GetAgentStatus
- README aktualisiert (Status-Doku, Events API, Dateistruktur)
This commit is contained in:
parent
5d7a3cbcc1
commit
c72b791796
40
README.md
40
README.md
@ -161,6 +161,42 @@ daemon -f -p /var/run/rmm-agent.pid -o /usr/local/rmm/rmm-agent.log \
|
||||
| GET | `/api/v1/agents/{id}/system` | Nur Systemdaten |
|
||||
| DELETE | `/api/v1/agents/{id}` | Agent loeschen |
|
||||
|
||||
### Agent-Status
|
||||
|
||||
Status wird automatisch aus `last_heartbeat` berechnet (kein DB-Feld):
|
||||
|
||||
| Status | Bedingung |
|
||||
|--------|-----------|
|
||||
| `online` | Heartbeat < 2 Minuten |
|
||||
| `stale` | Heartbeat < 5 Minuten |
|
||||
| `offline` | Heartbeat > 5 Minuten |
|
||||
| `unknown` | Kein Heartbeat vorhanden |
|
||||
|
||||
`GET /agents` liefert ein `status` Feld pro Agent, `GET /agents/{id}` ebenfalls.
|
||||
|
||||
### Agent-Events
|
||||
|
||||
| Methode | Endpoint | Beschreibung |
|
||||
|---------|----------|-------------|
|
||||
| GET | `/api/v1/agents/events` | Alle Events (`?limit=N&type=X`) |
|
||||
| GET | `/api/v1/agents/{id}/events` | Events eines Agents (`?limit=N&type=X`) |
|
||||
|
||||
Event-Typen: `online`, `offline`, `stale`, `connected`, `disconnected`
|
||||
|
||||
Events werden automatisch geschrieben bei:
|
||||
- WebSocket Connect/Disconnect (`connected`/`disconnected`)
|
||||
- Inaktivitaets-Monitor erkennt Statuswechsel (`online`/`stale`/`offline`)
|
||||
|
||||
```bash
|
||||
# Alle Events (letzte 10)
|
||||
curl -sk -H "X-API-Key: 01532e5a7c9e70bf2757df77a2f5b9b9" \
|
||||
"https://192.168.85.13:8443/api/v1/agents/events?limit=10"
|
||||
|
||||
# Nur Offline-Events eines Agents
|
||||
curl -sk -H "X-API-Key: 01532e5a7c9e70bf2757df77a2f5b9b9" \
|
||||
"https://192.168.85.13:8443/api/v1/agents/6beb8dfd38ecf03eab6c0f21b4ac7bee/events?type=offline"
|
||||
```
|
||||
|
||||
### WebSocket
|
||||
|
||||
| Methode | Endpoint | Beschreibung |
|
||||
@ -518,8 +554,10 @@ rmm/
|
||||
│ ├── models/
|
||||
│ │ ├── agent.go # Agent, Register/Heartbeat Structs
|
||||
│ │ └── system.go # SystemData, alle Collector-Structs
|
||||
│ ├── monitor/
|
||||
│ │ └── monitor.go # Inaktivitaets-Monitor (prueft Heartbeat, schreibt Events)
|
||||
│ ├── api/
|
||||
│ │ ├── handlers.go # REST API Handler
|
||||
│ │ ├── handlers.go # REST API Handler (inkl. Agent-Status + Events)
|
||||
│ │ ├── tunnel_proxy.go # Tunnel + Exec + Update REST Endpoints
|
||||
│ │ ├── backup.go # Config-Backup Endpoints (trigger, list, download, diff)
|
||||
│ │ ├── wireguard.go # WireGuard Peer API Endpoints
|
||||
|
||||
@ -6,6 +6,7 @@ import (
|
||||
"encoding/json"
|
||||
"log"
|
||||
"net/http"
|
||||
"strconv"
|
||||
"strings"
|
||||
"time"
|
||||
|
||||
@ -28,7 +29,9 @@ func (h *Handler) SetupRoutes(mux *http.ServeMux, hub *ws.Hub) {
|
||||
mux.HandleFunc("POST /api/v1/agent/register", h.registerAgent)
|
||||
mux.HandleFunc("POST /api/v1/agent/heartbeat", h.heartbeat)
|
||||
mux.HandleFunc("GET /api/v1/agents", h.listAgents)
|
||||
mux.HandleFunc("GET /api/v1/agents/events", h.getAllAgentEvents)
|
||||
mux.HandleFunc("GET /api/v1/agents/{id}", h.getAgent)
|
||||
mux.HandleFunc("GET /api/v1/agents/{id}/events", h.getAgentEvents)
|
||||
mux.HandleFunc("GET /api/v1/agents/{id}/system", h.getSystemData)
|
||||
mux.HandleFunc("DELETE /api/v1/agents/{id}", h.deleteAgent)
|
||||
|
||||
@ -151,7 +154,19 @@ func (h *Handler) listAgents(w http.ResponseWriter, r *http.Request) {
|
||||
if agents == nil {
|
||||
agents = []models.Agent{}
|
||||
}
|
||||
writeJSON(w, http.StatusOK, agents)
|
||||
|
||||
// Status fuer jeden Agent berechnen
|
||||
var response []models.AgentResponse
|
||||
for _, a := range agents {
|
||||
response = append(response, models.AgentResponse{
|
||||
Agent: a,
|
||||
Status: models.CalculateAgentStatus(a.LastHeartbeat),
|
||||
})
|
||||
}
|
||||
if response == nil {
|
||||
response = []models.AgentResponse{}
|
||||
}
|
||||
writeJSON(w, http.StatusOK, response)
|
||||
}
|
||||
|
||||
// GET /api/v1/agents/{id}
|
||||
@ -170,6 +185,7 @@ func (h *Handler) getAgent(w http.ResponseWriter, r *http.Request) {
|
||||
|
||||
result := map[string]interface{}{
|
||||
"agent": agent,
|
||||
"status": models.CalculateAgentStatus(agent.LastHeartbeat),
|
||||
"system_data": sysData,
|
||||
}
|
||||
writeJSON(w, http.StatusOK, result)
|
||||
@ -207,3 +223,42 @@ func (h *Handler) deleteAgent(w http.ResponseWriter, r *http.Request) {
|
||||
log.Printf("Agent geloescht: %s", id)
|
||||
writeJSON(w, http.StatusOK, map[string]string{"message": "Agent geloescht"})
|
||||
}
|
||||
|
||||
// GET /api/v1/agents/{id}/events
|
||||
func (h *Handler) getAgentEvents(w http.ResponseWriter, r *http.Request) {
|
||||
agentID := r.PathValue("id")
|
||||
eventType := r.URL.Query().Get("type")
|
||||
limit := 100
|
||||
if l := r.URL.Query().Get("limit"); l != "" {
|
||||
if n, err := strconv.Atoi(l); err == nil && n > 0 {
|
||||
limit = n
|
||||
}
|
||||
}
|
||||
|
||||
events, err := h.db.GetAgentEvents(agentID, eventType, limit)
|
||||
if err != nil {
|
||||
log.Printf("Fehler beim Laden der Events: %v", err)
|
||||
writeError(w, http.StatusInternalServerError, "Fehler beim Laden")
|
||||
return
|
||||
}
|
||||
writeJSON(w, http.StatusOK, events)
|
||||
}
|
||||
|
||||
// GET /api/v1/agents/events
|
||||
func (h *Handler) getAllAgentEvents(w http.ResponseWriter, r *http.Request) {
|
||||
eventType := r.URL.Query().Get("type")
|
||||
limit := 100
|
||||
if l := r.URL.Query().Get("limit"); l != "" {
|
||||
if n, err := strconv.Atoi(l); err == nil && n > 0 {
|
||||
limit = n
|
||||
}
|
||||
}
|
||||
|
||||
events, err := h.db.GetAllAgentEvents(eventType, limit)
|
||||
if err != nil {
|
||||
log.Printf("Fehler beim Laden der Events: %v", err)
|
||||
writeError(w, http.StatusInternalServerError, "Fehler beim Laden")
|
||||
return
|
||||
}
|
||||
writeJSON(w, http.StatusOK, events)
|
||||
}
|
||||
|
||||
@ -88,6 +88,15 @@ func (d *Database) migrate() error {
|
||||
|
||||
CREATE INDEX IF NOT EXISTS idx_config_backups_agent ON config_backups(agent_id, created_at DESC);
|
||||
CREATE INDEX IF NOT EXISTS idx_config_backups_hash ON config_backups(agent_id, hash);
|
||||
|
||||
CREATE TABLE IF NOT EXISTS agent_events (
|
||||
id SERIAL PRIMARY KEY,
|
||||
agent_id TEXT NOT NULL REFERENCES agents(id) ON DELETE CASCADE,
|
||||
event_type TEXT NOT NULL,
|
||||
message TEXT,
|
||||
created_at TIMESTAMPTZ NOT NULL DEFAULT NOW()
|
||||
);
|
||||
CREATE INDEX IF NOT EXISTS idx_agent_events_agent ON agent_events(agent_id, created_at DESC);
|
||||
`
|
||||
|
||||
if _, err := d.db.Exec(schema); err != nil {
|
||||
@ -631,3 +640,101 @@ func (d *Database) GetMetricsSummary(agentID, metric, bucket string, from, to ti
|
||||
}
|
||||
return results, rows.Err()
|
||||
}
|
||||
|
||||
// --- Agent Events Methoden ---
|
||||
|
||||
// InsertAgentEvent schreibt ein Event in die agent_events Tabelle
|
||||
func (d *Database) InsertAgentEvent(agentID, eventType, message string) error {
|
||||
_, err := d.db.Exec(
|
||||
"INSERT INTO agent_events (agent_id, event_type, message) VALUES ($1, $2, $3)",
|
||||
agentID, eventType, message,
|
||||
)
|
||||
if err != nil {
|
||||
log.Printf("Event schreiben fehlgeschlagen: %v", err)
|
||||
}
|
||||
return err
|
||||
}
|
||||
|
||||
// GetAgentEvents holt Events fuer einen bestimmten Agent
|
||||
func (d *Database) GetAgentEvents(agentID string, eventType string, limit int) ([]models.AgentEvent, error) {
|
||||
if limit <= 0 {
|
||||
limit = 100
|
||||
}
|
||||
|
||||
query := "SELECT id, agent_id, event_type, COALESCE(message, ''), created_at FROM agent_events WHERE agent_id = $1"
|
||||
args := []interface{}{agentID}
|
||||
|
||||
if eventType != "" {
|
||||
query += " AND event_type = $2"
|
||||
args = append(args, eventType)
|
||||
}
|
||||
|
||||
query += fmt.Sprintf(" ORDER BY created_at DESC LIMIT $%d", len(args)+1)
|
||||
args = append(args, limit)
|
||||
|
||||
rows, err := d.db.Query(query, args...)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
defer rows.Close()
|
||||
|
||||
var events []models.AgentEvent
|
||||
for rows.Next() {
|
||||
var e models.AgentEvent
|
||||
if err := rows.Scan(&e.ID, &e.AgentID, &e.EventType, &e.Message, &e.CreatedAt); err != nil {
|
||||
return nil, err
|
||||
}
|
||||
events = append(events, e)
|
||||
}
|
||||
if events == nil {
|
||||
events = []models.AgentEvent{}
|
||||
}
|
||||
return events, rows.Err()
|
||||
}
|
||||
|
||||
// GetAllAgentEvents holt Events fuer alle Agents
|
||||
func (d *Database) GetAllAgentEvents(eventType string, limit int) ([]models.AgentEvent, error) {
|
||||
if limit <= 0 {
|
||||
limit = 100
|
||||
}
|
||||
|
||||
query := "SELECT id, agent_id, event_type, COALESCE(message, ''), created_at FROM agent_events"
|
||||
args := []interface{}{}
|
||||
|
||||
if eventType != "" {
|
||||
query += " WHERE event_type = $1"
|
||||
args = append(args, eventType)
|
||||
}
|
||||
|
||||
query += fmt.Sprintf(" ORDER BY created_at DESC LIMIT $%d", len(args)+1)
|
||||
args = append(args, limit)
|
||||
|
||||
rows, err := d.db.Query(query, args...)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
defer rows.Close()
|
||||
|
||||
var events []models.AgentEvent
|
||||
for rows.Next() {
|
||||
var e models.AgentEvent
|
||||
if err := rows.Scan(&e.ID, &e.AgentID, &e.EventType, &e.Message, &e.CreatedAt); err != nil {
|
||||
return nil, err
|
||||
}
|
||||
events = append(events, e)
|
||||
}
|
||||
if events == nil {
|
||||
events = []models.AgentEvent{}
|
||||
}
|
||||
return events, rows.Err()
|
||||
}
|
||||
|
||||
// GetAgentStatus berechnet den Status eines Agents aus last_heartbeat
|
||||
func (d *Database) GetAgentStatus(agentID string) string {
|
||||
var lastHB sql.NullTime
|
||||
err := d.db.QueryRow("SELECT last_heartbeat FROM agents WHERE id = $1", agentID).Scan(&lastHB)
|
||||
if err != nil || !lastHB.Valid {
|
||||
return "unknown"
|
||||
}
|
||||
return models.CalculateAgentStatus(&lastHB.Time)
|
||||
}
|
||||
|
||||
@ -19,6 +19,7 @@ import (
|
||||
"github.com/cynfo/rmm-backend/api"
|
||||
"github.com/cynfo/rmm-backend/config"
|
||||
"github.com/cynfo/rmm-backend/db"
|
||||
"github.com/cynfo/rmm-backend/monitor"
|
||||
"github.com/cynfo/rmm-backend/ws"
|
||||
)
|
||||
|
||||
@ -47,8 +48,15 @@ func main() {
|
||||
|
||||
// WebSocket Hub starten
|
||||
hub := ws.NewHub()
|
||||
hub.SetEventCallback(func(agentID, eventType, message string) {
|
||||
database.InsertAgentEvent(agentID, eventType, message)
|
||||
})
|
||||
go hub.Run()
|
||||
|
||||
// Inaktivitaets-Monitor starten
|
||||
mon := monitor.New(database)
|
||||
go mon.Run()
|
||||
|
||||
// Router aufbauen
|
||||
mux := http.NewServeMux()
|
||||
handler := api.NewHandler(database)
|
||||
|
||||
@ -13,6 +13,37 @@ type Agent struct {
|
||||
LastHeartbeat *time.Time `json:"last_heartbeat,omitempty"`
|
||||
}
|
||||
|
||||
// AgentResponse erweitert Agent um den berechneten Status
|
||||
type AgentResponse struct {
|
||||
Agent
|
||||
Status string `json:"status"`
|
||||
}
|
||||
|
||||
// AgentEvent repraesentiert ein Event in der agent_events Tabelle
|
||||
type AgentEvent struct {
|
||||
ID int `json:"id"`
|
||||
AgentID string `json:"agent_id"`
|
||||
EventType string `json:"event_type"`
|
||||
Message string `json:"message,omitempty"`
|
||||
CreatedAt time.Time `json:"created_at"`
|
||||
}
|
||||
|
||||
// CalculateAgentStatus berechnet den Status aus dem letzten Heartbeat
|
||||
func CalculateAgentStatus(lastHeartbeat *time.Time) string {
|
||||
if lastHeartbeat == nil {
|
||||
return "unknown"
|
||||
}
|
||||
age := time.Since(*lastHeartbeat)
|
||||
switch {
|
||||
case age < 2*time.Minute:
|
||||
return "online"
|
||||
case age < 5*time.Minute:
|
||||
return "stale"
|
||||
default:
|
||||
return "offline"
|
||||
}
|
||||
}
|
||||
|
||||
// RegisterRequest - Agent-Registrierung
|
||||
type RegisterRequest struct {
|
||||
AgentID string `json:"agent_id,omitempty"`
|
||||
|
||||
77
backend/monitor/monitor.go
Normal file
77
backend/monitor/monitor.go
Normal file
@ -0,0 +1,77 @@
|
||||
package monitor
|
||||
|
||||
import (
|
||||
"log"
|
||||
"sync"
|
||||
"time"
|
||||
|
||||
"github.com/cynfo/rmm-backend/db"
|
||||
"github.com/cynfo/rmm-backend/models"
|
||||
)
|
||||
|
||||
// Monitor prueft periodisch den Agent-Status und schreibt Events bei Statuswechsel
|
||||
type Monitor struct {
|
||||
db *db.Database
|
||||
lastStatus map[string]string
|
||||
lastStatusMu sync.Mutex
|
||||
interval time.Duration
|
||||
}
|
||||
|
||||
// New erstellt einen neuen Monitor
|
||||
func New(database *db.Database) *Monitor {
|
||||
return &Monitor{
|
||||
db: database,
|
||||
lastStatus: make(map[string]string),
|
||||
interval: 60 * time.Second,
|
||||
}
|
||||
}
|
||||
|
||||
// Run startet den Monitor als Goroutine — blockiert, also mit go aufrufen
|
||||
func (m *Monitor) Run() {
|
||||
log.Println("Inaktivitaets-Monitor gestartet (Intervall: 60s)")
|
||||
|
||||
// Initialen Status laden
|
||||
m.checkAgents()
|
||||
|
||||
ticker := time.NewTicker(m.interval)
|
||||
defer ticker.Stop()
|
||||
|
||||
for range ticker.C {
|
||||
m.checkAgents()
|
||||
}
|
||||
}
|
||||
|
||||
func (m *Monitor) checkAgents() {
|
||||
agents, err := m.db.GetAgents()
|
||||
if err != nil {
|
||||
log.Printf("Monitor: Fehler beim Laden der Agents: %v", err)
|
||||
return
|
||||
}
|
||||
|
||||
m.lastStatusMu.Lock()
|
||||
defer m.lastStatusMu.Unlock()
|
||||
|
||||
for _, agent := range agents {
|
||||
newStatus := models.CalculateAgentStatus(agent.LastHeartbeat)
|
||||
oldStatus, known := m.lastStatus[agent.ID]
|
||||
|
||||
if !known {
|
||||
// Erster Check — Status merken, kein Event
|
||||
m.lastStatus[agent.ID] = newStatus
|
||||
log.Printf("Monitor: Agent %s (%s) initialer Status: %s", agent.Name, agent.ID, newStatus)
|
||||
continue
|
||||
}
|
||||
|
||||
if oldStatus == newStatus {
|
||||
continue
|
||||
}
|
||||
|
||||
// Statuswechsel erkannt
|
||||
log.Printf("Monitor: Agent %s (%s) Status-Wechsel: %s -> %s", agent.Name, agent.ID, oldStatus, newStatus)
|
||||
m.lastStatus[agent.ID] = newStatus
|
||||
|
||||
// Event schreiben
|
||||
message := "Status-Wechsel: " + oldStatus + " -> " + newStatus
|
||||
m.db.InsertAgentEvent(agent.ID, newStatus, message)
|
||||
}
|
||||
}
|
||||
@ -26,6 +26,9 @@ type Hub struct {
|
||||
// Response-Waiters: Command-ID -> Response-Channel
|
||||
responseWaiters map[string]chan map[string]interface{}
|
||||
waitersMux sync.RWMutex
|
||||
|
||||
// Event-Callback fuer DB-Events bei Connect/Disconnect
|
||||
eventCallback func(agentID, eventType, message string)
|
||||
}
|
||||
|
||||
func NewHub() *Hub {
|
||||
@ -41,6 +44,11 @@ func NewHub() *Hub {
|
||||
return hub
|
||||
}
|
||||
|
||||
// SetEventCallback setzt die Callback-Funktion fuer Connect/Disconnect Events
|
||||
func (h *Hub) SetEventCallback(cb func(agentID, eventType, message string)) {
|
||||
h.eventCallback = cb
|
||||
}
|
||||
|
||||
func (h *Hub) Run() {
|
||||
log.Println("WebSocket Hub startet...")
|
||||
|
||||
@ -79,6 +87,11 @@ func (h *Hub) registerAgent(agent *AgentConnection) {
|
||||
agent.LastSeen = time.Now()
|
||||
|
||||
log.Printf("Agent %s: WebSocket-Verbindung registriert", agent.AgentID)
|
||||
|
||||
// Event schreiben
|
||||
if h.eventCallback != nil {
|
||||
go h.eventCallback(agent.AgentID, "connected", "WebSocket-Verbindung hergestellt")
|
||||
}
|
||||
}
|
||||
|
||||
func (h *Hub) unregisterAgent(agent *AgentConnection) {
|
||||
@ -89,7 +102,12 @@ func (h *Hub) unregisterAgent(agent *AgentConnection) {
|
||||
delete(h.agents, agent.AgentID)
|
||||
existing.Connection.Close()
|
||||
log.Printf("Agent %s: WebSocket-Verbindung entfernt", agent.AgentID)
|
||||
|
||||
|
||||
// Event schreiben
|
||||
if h.eventCallback != nil {
|
||||
go h.eventCallback(agent.AgentID, "disconnected", "WebSocket-Verbindung getrennt")
|
||||
}
|
||||
|
||||
// Alle Tunnel für diesen Agent schließen
|
||||
h.tunnelManager.CloseAgentTunnels(agent.AgentID)
|
||||
}
|
||||
|
||||
Loading…
x
Reference in New Issue
Block a user