Zammad

Helpdesk orienté tickets. Email, Telegram, canaux sociaux, SLA, base de connaissances. OIDC natif.

• Projet original : https://zammad.com/
• Remplace : Zendesk, Freshdesk, Jira Service Desk
• Connexion (SSO) : À activer via l’interface admin — collez les valeurs OIDC_* depuis l’onglet Environment une fois.

Étapes de configuration

1. Cliquez Deploy. Le premier démarrage est lent (2-3 min) — Elasticsearch et les migrations Rails se lancent au démarrage.
2. Visitez votre domaine Zammad. Complétez l’assistant de configuration (compte admin, nom d’organisation).
3. (Optionnel) Activez Keycloak SSO : Settings -> Security -> Third-party authentication -> OpenID Connect -> collez OIDC_CLIENT_ID, OIDC_CLIENT_SECRET, OIDC_ISSUER_URL depuis l’onglet Environment. Validez.

Ressources : Zammad intègre Elasticsearch qui réserve ~1,5 GB de RAM. Prévoyez un VPS ≥4 GB si vous lancez Zammad avec d’autres apps.

Variables d’environnement

Ces valeurs se trouvent dans l’onglet Environment du compose Dokploy. Les secrets aléatoires sont générés automatiquement au premier semi du template — vous n’avez pas à les générer vous-même.

Variable	Valeur par défaut
ZAMMAD_HOSTNAME	help.yourdomain.com
DB_PASSWORD	valeur aléatoire auto-générée

Domaine

• Service et port : zammad-nginx:8080
• Nom d’hôte : help.yourdomain.com

Le nom d’hôte est attaché automatiquement au semi du template ; modifiez-le dans l’onglet Domains avant de cliquer Deploy si vous souhaitez autre chose.

Fichier compose

Pour référence — c’est ce que le template déploie. Ne collez ceci nulle part. Le compose est semé dans Dokploy automatiquement ; les ajustements côté client se font dans les onglets Environment et Domains (décrits plus haut), jamais dans le compose lui-même.

# Zammad -- ticket-first help desk / omnichannel support. Native OIDC
# support (configured in Settings -> Security -> Third-party
# authentication -> OIDC after first admin sign-in).
#
# Heavier than Chatwoot: bundles Elasticsearch (needs ~1.5 GB RAM), a
# dedicated scheduler, websocket process, nginx front-end, + the usual
# postgres/redis/memcached. Plan for a 4+ GB VPS tier.

x-zammad-env: &zammad-env
  POSTGRESQL_HOST: db
  POSTGRESQL_PORT: "5432"
  POSTGRESQL_DB: zammad_production
  POSTGRESQL_USER: zammad
  POSTGRESQL_PASS: ${DB_PASSWORD}
  REDIS_URL: redis://redis:6379
  MEMCACHE_SERVERS: memcached:11211
  ELASTICSEARCH_SCHEMA: http
  ELASTICSEARCH_HOST: elasticsearch
  ELASTICSEARCH_PORT: "9200"
  ELASTICSEARCH_NAMESPACE: zammad

services:
  zammad-init:
    image: zammad/zammad:7.0.1-0035
    restart: on-failure
    command: ["zammad-init"]
    environment: *zammad-env
    volumes:
      - zammad-storage:/opt/zammad/storage
    depends_on:
      db:
        condition: service_healthy
      elasticsearch:
        condition: service_healthy
    networks:
      - default

  zammad-railsserver:
    image: zammad/zammad:7.0.1-0035
    restart: unless-stopped
    command: ["zammad-railsserver"]
    environment: *zammad-env
    volumes:
      - zammad-storage:/opt/zammad/storage
    depends_on:
      - zammad-init
    labels:
      - "vps.auto-update=patch"
    networks:
      - default

  zammad-websocket:
    image: zammad/zammad:7.0.1-0035
    restart: unless-stopped
    command: ["zammad-websocket"]
    environment: *zammad-env
    depends_on:
      - zammad-init
    labels:
      - "vps.auto-update=patch"
    networks:
      - default

  zammad-scheduler:
    image: zammad/zammad:7.0.1-0035
    restart: unless-stopped
    command: ["zammad-scheduler"]
    environment: *zammad-env
    volumes:
      - zammad-storage:/opt/zammad/storage
    depends_on:
      - zammad-init
    labels:
      - "vps.auto-update=patch"
    networks:
      - default

  # Public-facing: fronts railsserver + websocket under a single vhost.
  zammad-nginx:
    image: zammad/zammad:7.0.1-0035
    restart: unless-stopped
    command: ["zammad-nginx"]
    environment:
      <<: *zammad-env
      NGINX_SERVER_NAME: ${ZAMMAD_HOSTNAME}
      NGINX_CLIENT_MAX_BODY_SIZE: 50M
    depends_on:
      - zammad-railsserver
      - zammad-websocket
    labels:
      - "vps.auth.mode=public"
      - "vps.auth.oidc=true"
      - "vps.auth.groups=staff"
      - "vps.auth.oidc.redirect_uris=https://${ZAMMAD_HOSTNAME}/auth/openid_connect/callback"
      - "vps.auth.oidc.scopes=openid email profile"
      - "vps.auto-update=patch"
    networks:
      dokploy-network:
        aliases:
          - zammad
      default: {}

  db:
    image: postgres:16.13-alpine
    restart: unless-stopped
    environment:
      POSTGRES_USER: zammad
      POSTGRES_PASSWORD: ${DB_PASSWORD}
      POSTGRES_DB: zammad_production
    volumes:
      - db-data:/var/lib/postgresql/data
    healthcheck:
      test: ["CMD-SHELL", "pg_isready -U zammad"]
      interval: 10s
      timeout: 5s
      retries: 5
    labels:
      - "vps.auto-update=patch"
    networks:
      - default

  redis:
    image: redis:8.6.3-alpine3.23
    restart: unless-stopped
    labels:
      - "vps.auto-update=patch"
    networks:
      - default

  memcached:
    image: memcached:1.6.41-alpine3.23
    restart: unless-stopped
    command: ["memcached", "-m", "256"]
    labels:
      - "vps.auto-update=patch"
    networks:
      - default

  elasticsearch:
    image: docker.elastic.co/elasticsearch/elasticsearch:8.19.15
    restart: unless-stopped
    environment:
      discovery.type: single-node
      xpack.security.enabled: "false"
      ES_JAVA_OPTS: "-Xms512m -Xmx1024m"
    volumes:
      - es-data:/usr/share/elasticsearch/data
    healthcheck:
      test: ["CMD-SHELL", "curl -sf http://localhost:9200/_cluster/health | grep -qE '\"status\":\"(green|yellow)\"'"]
      interval: 15s
      timeout: 10s
      retries: 10
      start_period: 60s
    labels:
      - "vps.auto-update=off"
    networks:
      - default

volumes:
  zammad-storage:
  db-data:
  es-data:

networks:
  dokploy-network:
    external: true

---

<- Retour au catalogue des applications pré-configurées