Postgres Migration not working - database is empty

[Poschinski]

Hi,
i am trying to host bracket in production using docker as it is described here.

This is what my configuration looks like:

services:
  bracket-frontend:
    image: ghcr.io/evroon/bracket-frontend
    container_name: bracket-frontend
    environment:
      NODE_ENV: "production"
      VITE_API_BASE_URL: "https://bracket-api.my-domain.de"
    networks:
      - proxy
    labels:
      - "traefik.enable=true"
      - "traefik.docker.network=proxy"
      - "traefik.http.routers.bracket-fe.rule=Host(`bracket.my-domain.de`)"
      - "traefik.http.routers.bracket-fe.entrypoints=websecure"
      - "traefik.http.routers.bracket-fe.tls.certresolver=letsencrypt"
      - "traefik.http.services.bracket-fe.loadbalancer.server.port=3000"
    restart: unless-stopped

  bracket-backend:
    image: ghcr.io/evroon/bracket-backend
    container_name: bracket-backend
    environment:
      ENVIRONMENT: "PRODUCTION"
      PG_DSN: "postgresql://user:password@postgres:5432/bracket_prod"
      CORS_ORIGINS: "https://bracket.my-domain.de"
      JWT_SECRET: "mysecret..."
    volumes:
      - ./backend/static:/app/static
    networks:
      - proxy
      - bracket_lan
    labels:
      - "traefik.enable=true"
      - "traefik.docker.network=proxy"
      - "traefik.http.routers.bracket-api.rule=Host(`bracket-api.my-domain.de`)"
      - "traefik.http.routers.bracket-api.entrypoints=websecure"
      - "traefik.http.routers.bracket-api.tls.certresolver=letsencrypt"
      - "traefik.http.services.bracket-api.loadbalancer.server.port=8400"
    restart: unless-stopped
    depends_on:
      - postgres

  postgres:
    image: postgres
    restart: always
    environment:
      POSTGRES_DB: bracket_prod
      POSTGRES_USER: user
      POSTGRES_PASSWORD: password
    volumes:
      - ./postgres:/var/lib/postgresql
    networks:
      - bracket_lan

networks:
  proxy:
    external: true
  bracket_lan:
    driver: bridge

The frontend and the postgres instance are starting with no errors.

When the backend is starting up, is says it starts the migrations - but this leads to an error and the backend container is restarting.

Here is a part of the logs from the backend and the database:

bracket-backend   | [2026-04-04 14:44:32,454] [bracket] [INFO] Current env: PRODUCTION
bracket-backend   | [2026-04-04 14:44:37 +0000] [18] [INFO] Started server process [18]
bracket-backend   | [2026-04-04 14:44:37 +0000] [18] [INFO] Waiting for application startup.
bracket-backend   | [2026-04-04 14:44:38,398] [bracket] [INFO] Running migrations
postgres-1        | 2026-04-04 14:44:38.589 UTC [87] ERROR:  index "ix_users_email" does not exist
postgres-1        | 2026-04-04 14:44:38.589 UTC [87] STATEMENT:
postgres-1        |     DROP INDEX ix_users_email
bracket-backend   | [2026-04-04 14:44:38 +0000] [18] [ERROR] Traceback (most recent call last):
bracket-backend   |   File "/app/.venv/lib/python3.14/site-packages/sqlalchemy/engine/base.py", line 1967, in _exec_single_context
bracket-backend   |     self.dialect.do_execute(
bracket-backend   |     ~~~~~~~~~~~~~~~~~~~~~~~^
bracket-backend   |         cursor, str_statement, effective_parameters, context
bracket-backend   |         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
bracket-backend   |     )
bracket-backend   |     ^
bracket-backend   |   File "/app/.venv/lib/python3.14/site-packages/sqlalchemy/engine/default.py", line 951, in do_execute
bracket-backend   |     cursor.execute(statement, parameters)
bracket-backend   |     ~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^
bracket-backend   | psycopg2.errors.UndefinedObject: index "ix_users_email" does not exist
... here comes the trace
bracket-backend   | [SQL:
bracket-backend   | DROP INDEX ix_users_email]
bracket-backend   | (Background on this error at: https://sqlalche.me/e/20/f405)
bracket-backend   |
bracket-backend   | [2026-04-04 14:44:38 +0000] [18] [ERROR] Application startup failed. Exiting.
bracket-backend   | [2026-04-04 14:44:38 +0000] [18] [INFO] Worker exiting (pid: 18)
bracket-backend   | [2026-04-04 14:44:39 +0000] [11] [ERROR] Worker (pid:18) exited with code 3
bracket-backend   | [2026-04-04 14:44:39 +0000] [11] [ERROR] Shutting down: Master
bracket-backend   | [2026-04-04 14:44:39 +0000] [11] [ERROR] Reason: Worker failed to boot.
bracket-backend exited with code 3 (restarting)

Does anybody have the same issue?
I dont really know how to fix this.

Thanks in advance.