Deploy django project and flower

[Hamza-abughazaleh]

trying to deploy a Django project with Flower each one on a separate server

deploy.yml

# Name of your application. Used to uniquely configure containers.
service: backend

# Name of the container image.
image: backend

env:
  secret:
   all secrets


# Deploy to these servers.
servers:
  web:
    hosts:
      - <first IP>
    cmd: ./start.sh web
    healthcheck:
      path: health
      port: 8000
      interval: 5s
    labels:
      traefik.enable: true
      traefik.http.routers.djangobase-web.rule: 'Host(`domain name`)'
  worker:
    hosts:
      - <first IP>
    cmd: ./start.sh worker
  scheduler:
    hosts:
      - <first IP>
    cmd: ./start.sh scheduler
  flower:
    labels:
      traefik.enable: true
      traefik.http.routers.flower-web.rule: 'Host(`domain name`)'
    hosts:
      - <second IP>
    cmd: ./start.sh flower
    healthcheck:
      path: /
      port: 8000
      interval: 5s

# Credentials for your image host.
registry:
  digital ocean
  username:
    - API-TOKEN 
  password:
    -API-TOKEN 


# Use a different ssh user than root
ssh:
  user: user-name


traefik:
  # Expose traefik on 8080
  host_port: 8080

docker file

# Use an official Python runtime as a parent image
FROM python:3.11-slim

# create the app user
RUN addgroup --system app && adduser --system --group app

# Set environment variables
ENV PYTHONDONTWRITEBYTECODE 1
ENV PYTHONUNBUFFERED 1
ENV DJANGO_SETTINGS_MODULE=app.settings

# Set the working directory in the container
WORKDIR /app
RUN mkdir /app/staticfiles /app/uploadedfiles

RUN apt update && apt upgrade -y

RUN apt update \
  && apt install -y --no-install-recommends build-essential libpq-dev pkg-config\
  && rm -rf /var/lib/apt/lists/*

RUN apt update && apt install -y --no-install-recommends curl && apt clean && rm -rf /var/lib/apt/lists/*

RUN apt update && apt install -y --no-install-recommends default-libmysqlclient-dev && rm -rf /var/lib/apt/lists/*
RUN apt update && apt install -y --no-install-recommends postgresql-client && rm -rf /var/lib/apt/lists/*

# Copy the requirements file into the container at /app/
COPY requirements.txt /app/

# Install any needed packages specified in requirements.txt
RUN pip install --no-cache-dir -r requirements.txt

# Copy the current directory contents into the container at /app/
COPY code files

# Copy start script into the container
COPY bin/start.sh /app/start.sh

# Make the start script executable
RUN chmod +x /app/start.sh

RUN chown -R app:app /app
# change to the app user
USER app

EXPOSE 8000

CMD ["/app/start.sh", "web"]

start.sh file

#!/bin/bash

if [ "$1" == "web" ]; then
  # Wait for PostgreSQL to become available
  while ! pg_isready -h "${POSTGRES_DB_HOST}" -p "${POSTGRES_DB_PORT}" -U "${POSTGRES_DB_USER}"; do
      echo "Retrying pg_isready"
      sleep 1
  done
  set +x
  echo "PostgreSQL started."

  python manage.py collectstatic --noinput

  # Apply database migrations
  echo "Applying database migrations..."
  python manage.py migrate

  # Start the Django development server
  echo "Starting Django development server..."
  gunicorn --bind 0.0.0.0:8000 --timeout 120 app.wsgi:application

elif [ "$1" == "worker" ]; then
  shift
  celery -A app worker -E -l WARNING -n appCelery

elif [ "$1" == "scheduler" ]; then
  shift
  celery -A app beat -l WARNING

elif [ "$1" == "flower" ]; then
  shift
  celery -A app flower --port=8000
fi

The Django and Flower works on port 8000
when I access the first IP with port 8080 it's working fine but when I access the second IP with port 8080 it's not running the flower UI
Does anyone know how I can fix this issue?

[aalhour]

Hello guys 👋

Any update on this? We've been stuck with this issue for a while, tried to get help from the Discord server but no one helped us. Is it possible for us to deploy two web services on two different IPs with Kamal? If that's not possible we'd like to know as soon as possible to move on to an alternative tool.

Regards,
Ahmad

[aalhour]

@dhh, is there a specific person we can tag to an answer on this quickly?

[nickhammond]

@aalhour Kamal by default only enables traefik on the web role, if you want it on another server role you'll need to add traefik: true within the base level of that role. Your code snippet looks like it's currently nested under labels and slightly different.

https://kamal-deploy.org/docs/configuration#using-different-roles-for-servers

[Hamza-abughazaleh]

@nickhammond
cc @aalhour
We tried two IPs and the flower service is failing (container not healthy)

this the error that return from the kamal deploy

  Deploy lock was not released
  Finished all in 83.2 seconds
  ERROR (Kamal::Cli::Healthcheck::Poller::HealthcheckError): Exception while executing on host <second IP>: container not ready (unhealthy)

deploy.yaml

service: backend

# Name of the container image.
image: backend

env:
  secret:
   all secrets


# Deploy to these servers.
servers:
  web:
    hosts:
      - <first IP>
    cmd: ./start.sh web
    healthcheck:
      path: health
      port: 8000
      interval: 5s
  worker:
    hosts:
      - <first IP>
    cmd: ./start.sh worker
  scheduler:
    hosts:
      - <first IP>
    cmd: ./start.sh scheduler
  flower:
    traefik: true
    hosts:
      - <second IP>
    cmd: ./start.sh flower
    healthcheck:
      path: /
      port: 8000
      interval: 5s

# Credentials for your image host.
registry:
  digital ocean
  username:
    - API-TOKEN 
  password:
    -API-TOKEN 


# Use a different ssh user than root
ssh:
  user: user-name


traefik:
  # Expose traefik on 8080
  host_port: 8080

We tried hosting them on the same IP and both containers became healthy but traefik took flower and overridden the entrypoint of web with flower, meaning: accessing the IP over HTTP returns the flower service and not the web service.
File for two IPs

[nickhammond]

If you have them on the same machine then you're going to need to add a routing rule to tell traefik which traffic should go to each service. If you don't have a routing rule in place it's just going to route traffic to the first service it knows about.

Ideally, you point a domain or subdomain at your IP and set a routing rule to map to each service, your initial example had this in there, not seeing it in this example though.

[aalhour]

Thanks, but how can we host them on different machines? We followed your suggestion and the docs and it didn't work. Meaning, using the same deploy.yaml file, we want to host two web endpoints on two different machines with two different IP addresses.

[Hamza-abughazaleh]

@nickhammond Thanks, it works :)

[Hamza-abughazaleh]

@nickhammond we added these logic

labels:
  traefik.http.routers.djangobase-web.rule: Host(`domain name `)
  traefik.http.routers.flowerbase-web-flower.rule: Host(`domain name`)

Is it correct or not?

[nickhammond]

@Hamza-abughazaleh Yep, as long as those are nested within each server definition.