# 12 — Monitoring et alertes

---

## Stack de monitoring

```
Prometheus    ← Scrape des métriques (pods, nodes, app)
Grafana       ← Dashboards visuels
Loki          ← Agrégation des logs (NestJS pino)
Alertmanager  ← Envoi alertes (email, Slack)
Uptime Kuma   ← Monitoring externe HTTP (health checks)
```

---

## Installation du kube-prometheus-stack

La stack la plus complète, déployée avec Helm :

```bash
# Ajouter le repo
helm repo add prometheus-community https://prometheus-community.github.io/helm-charts
helm repo update

# Créer le namespace monitoring
kubectl create namespace monitoring

# Installer kube-prometheus-stack
helm install prometheus prometheus-community/kube-prometheus-stack \
  --namespace monitoring \
  --version 65.3.1 \
  --set grafana.adminPassword="<MOT_DE_PASSE_FORT>" \
  --set grafana.persistence.enabled=true \
  --set grafana.persistence.size=2Gi \
  --set grafana.persistence.storageClassName=hcloud-volumes \
  --set prometheus.prometheusSpec.retention=7d \
  --set prometheus.prometheusSpec.storageSpec.volumeClaimTemplate.spec.storageClassName=hcloud-volumes \
  --set prometheus.prometheusSpec.storageSpec.volumeClaimTemplate.spec.resources.requests.storage=10Gi \
  --set alertmanager.alertmanagerSpec.storage.volumeClaimTemplate.spec.storageClassName=hcloud-volumes \
  --set alertmanager.alertmanagerSpec.storage.volumeClaimTemplate.spec.resources.requests.storage=2Gi \
  --set prometheusOperator.resources.requests.cpu=50m \
  --set prometheusOperator.resources.requests.memory=128Mi \
  --set prometheus.prometheusSpec.resources.requests.cpu=100m \
  --set prometheus.prometheusSpec.resources.requests.memory=512Mi \
  --set grafana.resources.requests.cpu=50m \
  --set grafana.resources.requests.memory=128Mi

# Attendre que tout soit Running
kubectl rollout status deployment/prometheus-grafana -n monitoring --timeout=300s
kubectl get pods -n monitoring
```

---

## Exposer Grafana via Ingress

```bash
cat > /tmp/grafana-ingress.yaml << 'EOF'
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: grafana
  namespace: monitoring
  annotations:
    cert-manager.io/cluster-issuer: "letsencrypt-prod"
    traefik.ingress.kubernetes.io/router.entrypoints: "websecure"
    traefik.ingress.kubernetes.io/router.tls: "true"
    # Restreindre aux IPs de l'équipe
    traefik.ingress.kubernetes.io/router.middlewares: "monitoring-ipwhitelist@kubernetescrd"
spec:
  ingressClassName: traefik
  tls:
  - hosts:
    - monitoring.xpeditis.com
    secretName: monitoring-tls
  rules:
  - host: monitoring.xpeditis.com
    http:
      paths:
      - path: /
        pathType: Prefix
        backend:
          service:
            name: prometheus-grafana
            port:
              number: 80
---
# IP Whitelist pour Grafana (votre équipe seulement)
apiVersion: traefik.io/v1alpha1
kind: Middleware
metadata:
  name: ipwhitelist
  namespace: monitoring
spec:
  ipWhiteList:
    sourceRange:
      - "<VOTRE_IP>/32"
      - "10.0.0.0/16"    # Réseau interne Hetzner
EOF

kubectl apply -f /tmp/grafana-ingress.yaml
```

---

## Installation de Loki (agrégation des logs)

```bash
helm repo add grafana https://grafana.github.io/helm-charts
helm repo update

helm install loki grafana/loki-stack \
  --namespace monitoring \
  --set loki.persistence.enabled=true \
  --set loki.persistence.size=5Gi \
  --set loki.persistence.storageClassName=hcloud-volumes \
  --set promtail.enabled=true \
  --set loki.config.limits_config.retention_period=7d \
  --set grafana.enabled=false    # On utilise le Grafana déjà installé

# Ajouter Loki comme datasource dans Grafana
# Grafana → Data Sources → Add → Loki
# URL: http://loki:3100
```

---

## Configuration des alertes

### Alertes Xpeditis spécifiques

```bash
cat > /tmp/xpeditis-alerts.yaml << 'EOF'
apiVersion: monitoring.coreos.com/v1
kind: PrometheusRule
metadata:
  name: xpeditis-alerts
  namespace: xpeditis-prod
  labels:
    release: prometheus
spec:
  groups:
  - name: xpeditis.backend
    interval: 30s
    rules:

    # Backend down
    - alert: XpeditisBackendDown
      expr: up{job="xpeditis-backend"} == 0
      for: 1m
      labels:
        severity: critical
      annotations:
        summary: "Backend Xpeditis indisponible"
        description: "Aucun pod backend ne répond depuis 1 minute."

    # Trop peu de replicas
    - alert: XpeditisBackendLowReplicas
      expr: kube_deployment_status_replicas_available{deployment="xpeditis-backend",namespace="xpeditis-prod"} < 1
      for: 2m
      labels:
        severity: critical
      annotations:
        summary: "Moins d'1 replica backend disponible"

    # CPU élevé (déclenchement autoscaling probable)
    - alert: XpeditisHighCPU
      expr: |
        sum(rate(container_cpu_usage_seconds_total{
          namespace="xpeditis-prod",
          container="backend"
        }[5m])) by (pod) > 0.8
      for: 5m
      labels:
        severity: warning
      annotations:
        summary: "CPU élevé sur pod {{ $labels.pod }}"
        description: "Utilisation CPU > 80% depuis 5 minutes."

    # Mémoire élevée
    - alert: XpeditisHighMemory
      expr: |
        container_memory_usage_bytes{
          namespace="xpeditis-prod",
          container="backend"
        } / container_spec_memory_limit_bytes{
          namespace="xpeditis-prod",
          container="backend"
        } > 0.85
      for: 5m
      labels:
        severity: warning
      annotations:
        summary: "Mémoire élevée sur pod {{ $labels.pod }}"

    # Taux d'erreur HTTP élevé
    - alert: XpeditisHighErrorRate
      expr: |
        sum(rate(traefik_service_requests_total{
          service=~"xpeditis-prod-xpeditis-backend.*",
          code=~"5.."
        }[5m])) /
        sum(rate(traefik_service_requests_total{
          service=~"xpeditis-prod-xpeditis-backend.*"
        }[5m])) > 0.05
      for: 2m
      labels:
        severity: warning
      annotations:
        summary: "Taux d'erreur 5xx > 5% sur l'API backend"

    # Pods en CrashLoopBackOff
    - alert: XpeditisPodCrashLooping
      expr: |
        increase(kube_pod_container_status_restarts_total{
          namespace="xpeditis-prod"
        }[1h]) > 5
      labels:
        severity: critical
      annotations:
        summary: "Pod {{ $labels.pod }} redémarre trop souvent"

  - name: xpeditis.database
    rules:
    # Pas d'alerte directe sur Neon (managed) — uniquement si self-hosted

  - name: xpeditis.redis
    rules:
    # Redis mémoire élevée
    - alert: RedisHighMemory
      expr: |
        redis_memory_used_bytes /
        redis_memory_max_bytes > 0.85
      for: 5m
      labels:
        severity: warning
      annotations:
        summary: "Redis utilise > 85% de sa mémoire"
EOF

kubectl apply -f /tmp/xpeditis-alerts.yaml
```

### Configuration Alertmanager (Slack)

```bash
cat > /tmp/alertmanager-config.yaml << 'EOF'
apiVersion: v1
kind: Secret
metadata:
  name: alertmanager-prometheus-kube-prometheus-alertmanager
  namespace: monitoring
stringData:
  alertmanager.yaml: |
    global:
      resolve_timeout: 5m
      slack_api_url: '<SLACK_WEBHOOK_URL>'

    route:
      group_by: ['alertname', 'namespace']
      group_wait: 10s
      group_interval: 10m
      repeat_interval: 12h
      receiver: 'slack-notifications'
      routes:
      - match:
          severity: critical
        receiver: 'slack-critical'
      - match:
          severity: warning
        receiver: 'slack-notifications'

    receivers:
    - name: 'slack-notifications'
      slack_configs:
      - channel: '#xpeditis-monitoring'
        icon_url: https://avatars.githubusercontent.com/u/3380462
        title: '{{ template "slack.default.title" . }}'
        text: '{{ template "slack.default.text" . }}'
        send_resolved: true

    - name: 'slack-critical'
      slack_configs:
      - channel: '#xpeditis-alerts-critiques'
        color: 'danger'
        title: '🚨 ALERTE CRITIQUE : {{ .CommonAnnotations.summary }}'
        text: '{{ .CommonAnnotations.description }}'
        send_resolved: true
EOF

kubectl apply -f /tmp/alertmanager-config.yaml
```

---

## Dashboards Grafana recommandés

Importez ces dashboards depuis grafana.com (ID à entrer dans Grafana → Import) :

| Dashboard | ID | Usage |
|---|---|---|
| Kubernetes Cluster Overview | 6417 | Vue d'ensemble cluster |
| Kubernetes Deployments | 8588 | Détail des deployments |
| Node Exporter Full | 1860 | Métriques système des nœuds |
| Loki & Promtail | 12611 | Logs agrégés |
| Traefik 2 | 4475 | Métriques ingress/requêtes |

```bash
# Dans Grafana (https://monitoring.xpeditis.com)
# → + → Import
# → Entrer l'ID et cliquer "Load"
# → Sélectionner la datasource Prometheus
# → Import
```

---

## Uptime Kuma (monitoring externe)

Uptime Kuma monitore vos endpoints depuis l'extérieur du cluster, indépendamment de Prometheus :

```bash
# Déployer Uptime Kuma dans le cluster
cat > /tmp/uptime-kuma.yaml << 'EOF'
apiVersion: apps/v1
kind: Deployment
metadata:
  name: uptime-kuma
  namespace: monitoring
spec:
  replicas: 1
  selector:
    matchLabels:
      app: uptime-kuma
  template:
    metadata:
      labels:
        app: uptime-kuma
    spec:
      containers:
      - name: uptime-kuma
        image: louislam/uptime-kuma:1
        ports:
        - containerPort: 3001
        volumeMounts:
        - name: data
          mountPath: /app/data
        resources:
          requests:
            cpu: 50m
            memory: 128Mi
      volumes:
      - name: data
        persistentVolumeClaim:
          claimName: uptime-kuma-pvc
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: uptime-kuma-pvc
  namespace: monitoring
spec:
  accessModes: [ReadWriteOnce]
  storageClassName: hcloud-volumes
  resources:
    requests:
      storage: 1Gi
---
apiVersion: v1
kind: Service
metadata:
  name: uptime-kuma
  namespace: monitoring
spec:
  selector:
    app: uptime-kuma
  ports:
  - port: 3001
    targetPort: 3001
EOF

kubectl apply -f /tmp/uptime-kuma.yaml
```

Monitors à configurer dans Uptime Kuma :

| Monitor | URL | Intervalle |
|---|---|---|
| API Health | `https://api.xpeditis.com/api/v1/health` | 1 min |
| Frontend | `https://app.xpeditis.com/` | 1 min |
| API Login | `POST https://api.xpeditis.com/api/v1/auth/login` | 5 min |

---

## Commandes de monitoring rapides

```bash
# Top des pods par consommation CPU/RAM
kubectl top pods -n xpeditis-prod --sort-by=cpu

# Événements récents du namespace
kubectl get events -n xpeditis-prod --sort-by='.lastTimestamp' | tail -20

# Logs backend en temps réel (tous les pods)
stern xpeditis-backend -n xpeditis-prod

# Logs d'erreurs uniquement
kubectl logs -l app=xpeditis-backend -n xpeditis-prod --since=1h | grep -i error

# Status des HPAs
kubectl get hpa -n xpeditis-prod

# Métriques des nœuds
kubectl top nodes
```