Huginn

Huginn is available via two different Docker image strategies depending on your deployment needs:

1. huginn/huginn (Multiple Process Image): Runs all Huginn processes within a single container. If no external database is linked, it automatically starts an internal MySQL instance. This is ideal for quick testing and prototyping, but it is more difficult to maintain and back up.

2. huginn/huginn-single-process (Single Process Image): Follows the one-process-per-container pattern. This requires running multiple containers (at minimum one for the application server and one for the threaded background worker). This approach is more scalable and allows you to run multiple background workers in separate containers to improve performance.

For production environments where you need to scale components individually, use the huginn/huginn-single-process image. This allows you to run each Huginn process (such as the web server, workers, etc.) in its own dedicated container.

Detailed documentation for the single-process architecture can be found in the repository's docker/single-process directory.

A DelayedJob worker is a separate process that executes Huginn Agents (e.g., polling services, fetching websites).

Scaling Logic:

• One worker can perform exactly one check at a time.
• Example: If you have 60 Agents checking websites every minute, and each check takes 1 second, a single worker is sufficient.
• When to add workers: If you have a high volume of Agents or if the services/websites being polled are slow or unreliable, you should run additional workers.

Changes to the .env or Procfile must be performed directly on the server. After modifying these files, you must export the init scripts and restart Huginn to apply the changes.

1. Navigate to the Huginn directory: cd /home/huginn/huginn
2. Edit the desired file (e.g., Procfile or .env) using the huginn user.
3. Export init scripts and restart: sudo rake production:export

cd /home/huginn/huginn
# Edit files as needed
sudo -u huginn -H editor Procfile
sudo -u huginn -H editor .env
# Export init scripts and restart huginn
sudo rake production:export

You can run a local instance of Huginn using Docker for testing or development purposes. Note that Docker is not recommended for production use.

For detailed instructions, refer to the Docker installation guide.

# See docker/install.md for specific commands

To use PostgreSQL, install the server and libpq-dev. You will need to create a PostgreSQL user and a database named huginn_production owned by that user.

# Install packages
sudo apt-get install -y postgresql libpq-dev

# Create Huginn user (will prompt for password)
sudo -u postgres -H createuser -P huginn

# Create the production database
sudo -u postgres -H createdb -O huginn -T template0 huginn_production

# Verify connection
sudo -u huginn psql -h localhost -W huginn_production

Huginn uses foreman to manage processes defined in the Procfile.

1. Edit Procfile: You can modify the Procfile to change process layouts (e.g., enabling separate schedule, twitter, or dj workers). Ensure there are no leading spaces before web: or jobs:.
2. Export Scripts: Use the production export task to generate the init scripts.
3. Re-exporting: You must re-run the export command every time you modify .env, Procfile, or config/puma.rb.

Custom Log Directory: If you use a directory other than /home/huginn/huginn/, update the runit logfile location in lib/tasks/production.rake (e.g., run('foreman export runit -a huginn -l /opt/huginn/log /etc/service')).

# Export the init scripts
sudo bundle exec rake production:export

You can deploy Huginn to OpenShift Online using provided JSON templates. Use the following commands to create a new app with either MySQL or PostgreSQL:

For MySQL:

oc new-app -f https://raw.githubusercontent.com/huginn/huginn/master/openshift/templates/huginn-mysql.json

For PostgreSQL:

oc new-app -f https://raw.githubusercontent.com/huginn/huginn/master/openshift/templates/huginn-postgresql.json

To use Capistrano for deployments, add the following configuration keys to your local .env file. This tells Capistrano where to deploy, which user to use, and which repository to pull from.

CAPISTRANO_DEPLOY_SERVER=<IP or FQDN of your server>
CAPISTRANO_DEPLOY_USER=huginn
CAPISTRANO_DEPLOY_REPO_URL=https://github.com/huginn/huginn.git