Need help with configuration for Docker Swarm + Traefik

[florentin68]

Hello,
Thank you for this amazing project.
I'm using Docker Swarm with multiple services.
I configured such as mentionned in SablierApp documentation (Traefik plugin with 1 middleware to test, groupname=default, Sablier service with sablier.yaml config file, labels on services I want to scale to zero, Traefik whoami for example to test).
When I try to access whoami.domain.com I get this error:

{
    "availableGroups": null,
    "detail": "The group you requested does not exist. It is possible that the group has not been scanned yet.",
    "error": "group default not found",
    "instance": "/api/strategies/dynamic?display_name=Loading...\u0026group=default\u0026refresh_frequency=2s\u0026session_duration=30m\u0026show_details=true\u0026theme=hacker-terminal",
    "requestGroup": "home",
    "status": 404,
    "title": "Group not found",
    "type": "https://sablierapp.dev/#/errors?id=group-not-found"
}

What did I miss?

[uldiseihenbergs]

Post your config. you should add group name to your service(with labels) and it should match with group name in middleware

[florentin68]

Here are config files extractions:

compose.yml

  whoami:
    image: traefik/whoami
    networks:
      - t2_proxy
    deploy:
      replicas: 0 # [sablier]
      labels:
        ## Traefik
        - traefik.enable=true
        - traefik.swarm.lbswarm=true
        ## HTTP Routers
        - 'traefik.http.routers.whoami.entrypoints=https'
        - 'traefik.http.routers.whoami.rule=Host(`whoami.${DOMAINNAME}`)'
        ## HTTP Middlewares
        - 'traefik.http.routers.whoami.middlewares=my-sablier@file'
        ## HTTP Services
        - 'traefik.http.routers.whoami.service=whoami'
        - 'traefik.http.services.whoami.loadbalancer.server.port=80'
        ## Diun
        - diun.enable=true
        - diun.watch_repo=false
        ## Wud
        - wud.watch=true
        ## Sablier
        - sablier.enable=true
        - sablier.group=default
      restart_policy:
        condition: any
        delay: 5s
        max_attempts: 3
        window: 120s

  # Sablier - On-demand services
  sablier:
    image: ghcr.io/sablierapp/sablier
    networks:
      - t2_proxy
    volumes:
      - /var/run/docker.sock:/var/run/docker.sock
      - $DOCKERDIR/volumes/sablier/sablier.yml:/etc/sablier/sablier.yml
    deploy:
      labels:
        ## Traefik
        - traefik.enable=false
        ## Diun
        - diun.enable=true
        - diun.watch_repo=false
        ## Wud
        - wud.watch=true
      placement:
        constraints: [node.role == manager]
      update_config:
        parallelism: 2
        delay: 10s
      restart_policy:
        condition: any
        delay: 5s
        max_attempts: 3
        window: 120s

traefik-middleware.yml

    my-sablier:
      plugin:
        sablier:
          dynamic:
            displayName: "Loading..."
            refreshFrequency: 2s
            showDetails: "true"
            theme: hacker-terminal
          group: default
          #names: general_whoami,dashy,heimdall,phpmyadmin,pgadmin,ldapadmin,grafana#,nginx
          sablierUrl: http://sablier:10000
          sessionDuration: 30m

[uldiseihenbergs]

post your /etc/sablier/sablier.yml too.
and if you use group, sablier will start and stop all services at once..
if you set sublier logging to trace, what it contains at sablier service startup? it should write info about groups and services into it..

[florentin68]

Sorry, I missed it:
sablier.yml

provider:
  # Provider to use to manage containers (docker, swarm, kubernetes)
  name: swarm
server:
  # The server port to use
  port: 10000
  # The base path for the API
  base-path: /
storage:
  # File path to save the state (default stateless)
  file:
sessions:
  # The default session duration (default 5m)
  default-duration: 5m
  # The expiration checking interval.
  # Higher duration gives less stress on CPU.
  # If you only use sessions of 1h, setting this to 5m is a good trade-off.
  expiration-interval: 20s
logging:
  level: info
strategy:
  dynamic:
    # Custom themes folder, will load all .html files recursively (default empty)
    custom-themes-path:
    # Show instances details by default in waiting UI
    show-details-by-default: false
    # Default theme used for dynamic strategy (default "hacker-terminal")
    default-theme: hacker-terminal
    # Default refresh frequency in the HTML page for dynamic strategy
    default-refresh-frequency: 5s
  blocking:
    # Default timeout used for blocking strategy (default 1m)
    default-timeout: 1m

With trace loglevel instead of info I get this at startup:

general_sablier.1.mghd9m9l9h4j@photon1-fm    | 7:57PM INF sablier/sablier.go:25 running Sablier version (version=1.9.0, branch=HEAD, revision=31464c38b5859738e4b1c9f1962de833d47db001)
general_sablier.1.mghd9m9l9h4j@photon1-fm    | 7:57PM INF docker/docker.go:28 connection established with docker provider=docker version=27.3.1 api_version=1.47
general_sablier.1.mghd9m9l9h4j@photon1-fm    | 7:57PM INF theme/parse.go:12 theme found path=embedded/ghost.html
general_sablier.1.mghd9m9l9h4j@photon1-fm    | 7:57PM INF theme/parse.go:19 successfully added theme path=embedded/ghost.html
general_sablier.1.mghd9m9l9h4j@photon1-fm    | 7:57PM INF theme/parse.go:12 theme found path=embedded/hacker-terminal.html
general_sablier.1.mghd9m9l9h4j@photon1-fm    | 7:57PM INF theme/parse.go:19 successfully added theme path=embedded/hacker-terminal.html
general_sablier.1.mghd9m9l9h4j@photon1-fm    | 7:57PM INF theme/parse.go:12 theme found path=embedded/matrix.html
general_sablier.1.mghd9m9l9h4j@photon1-fm    | 7:57PM INF theme/parse.go:19 successfully added theme path=embedded/matrix.html
general_sablier.1.mghd9m9l9h4j@photon1-fm    | 7:57PM INF theme/parse.go:12 theme found path=embedded/shuffle.html
general_sablier.1.mghd9m9l9h4j@photon1-fm    | 7:57PM INF theme/parse.go:19 successfully added theme path=embedded/shuffle.html
general_sablier.1.mghd9m9l9h4j@photon1-fm    | 7:57PM INF server/server.go:43 starting  listen=:10000 startup=76.924µs mode=release

Provider = docker, not swarm. It seems the config file isn't parsed, no?

File attributes in host:

root@photon1-fm [ /share/homes/docker/volumes/sablier ]# ls -al
total 12
drwxr-x---  2 root root 4096 May 26 21:55 .
drwxr-xr-x 63 root adm  4096 May 25 21:35 ..
-rw-r-----  1 root root 1076 May 26 21:55 sablier.yml

[uldiseihenbergs]

looks like sablier can't read your config file. as a test, give 777 to sablier.yml on host. and then look inside docker for this file, and check permissions

[florentin68]

I didn't manage to connect to container with docker exec -it sablier sh or bash

I switched to 777 but it's not better. I may have a syntax problem in my compose file but I don't see it.

[uldiseihenbergs]

you can add configuration with your compose file. i'm running it this way, for example:

sablier:
image: sablierapp/sablier:1.9.0
command:
- start
- --provider.name=docker_swarm # or swarm
- --provider.auto-stop-on-startup=false
- --strategy.blocking.default-refresh-frequency=2s
- --logging.level=trace

[florentin68]

With the command lines in compose file it is a little bit better: containers in group general are starting when needed.
How can I manage each container on its own? Is it possible using only 1 middleware config in Traefik?
Log level = trace seems not recognized...

2025/05/30 09:23:03 WARN invalid log level, defaulting to info level=trace
9:23AM INF sablier/sablier.go:25 running Sablier version (version=1.9.0, branch=HEAD, revision=31464c38b5859738e4b1c9f1962de833d47db001)
9:23AM INF dockerswarm/docker_swarm.go:34 connection established with docker swarm provider=swarm version=27.3.1 api_version=1.47
9:23AM INF sablier/sablier.go:44 set groups old=map[] new="map[default:[general_dashy general_zabbix-web general_heimdall general_pgadmin general_teleport general_grafana general_rediscommander general_phpmyadmin general_whoami]]" diff="  map[string][]string{\n+ \t\"default\": {\n+ \t\t\"general_dashy\", \"general_zabbix-web\", \"general_heimdall\", \"general_pgadmin\",\n+ \t\t\"general_teleport\", \"general_grafana\", \"general_rediscommander\",\n+ \t\t\"general_phpmyadmin\", \"general_whoami\",\n+ \t},\n  }\n"
9:23AM INF theme/parse.go:12 theme found path=embedded/ghost.html
9:23AM INF theme/parse.go:19 successfully added theme path=embedded/ghost.html
9:23AM INF theme/parse.go:12 theme found path=embedded/hacker-terminal.html
9:23AM INF theme/parse.go:19 successfully added theme path=embedded/hacker-terminal.html
9:23AM INF theme/parse.go:12 theme found path=embedded/matrix.html
9:23AM INF theme/parse.go:19 successfully added theme path=embedded/matrix.html
9:23AM INF theme/parse.go:12 theme found path=embedded/shuffle.html
9:23AM INF theme/parse.go:19 successfully added theme path=embedded/shuffle.html
9:23AM INF server/server.go:43 starting  listen=:10000 startup=73.869µs mode=release
9:57AM INF slog-gin@v1.14.1/middleware.go:258 Incoming request request.time="2025-05-30 09:57:53.635497861 +0000 UTC" request.method=GET request.host=sablier:10000 request.path=/api/strategies/dynamic request.query="display_name=Loading...&group=default&refresh_frequency=2s&session_duration=30m&show_details=true&theme=hacker-terminal" request.params=map[] request.route=/api/strategies/dynamic request.ip=192.168.90.12 request.referer="" request.length=0 response.time="2025-05-30 09:57:53.666664764 +0000 UTC" response.latency=31.166912ms response.status=200 response.length=6508 id=482dcdd4-c376-4a1e-8a37-21a211160fde

[uldiseihenbergs]

don't use groups then, but use names instead in middleware. and yes - one midleware for each container then.
according to loglevel - first line in log about "WARN" - use lowercase?

[florentin68]

Ok great, I fixed everything by creating middleware labels for each service to scale to 0, each associated to its own group.
I still don't understand why config file isn't read and what log levels can be set... but it's running.

Thanks for helping @uldiseihenbergs.

[uldiseihenbergs]

i'm using this in my docker-compose.yml, and log level works as expected. in info level you see all reguests to all contaiters - too much log.

command:
- start
- --provider.name=docker_swarm # or swarm
- --provider.auto-stop-on-startup=false
- --strategy.blocking.default-refresh-frequency=2s
- --logging.level=warn