All configuration is managed through environment variables. Copy the sample file to get started:
cp contrib/.env.sample .env
Edit .env in the project root before starting services. When running with Docker Compose, variables are loaded automatically via the env_file directive in docker-compose.yml.
Django settings Variable Type Description ENVIRONMENTstrRuntime environment label, e.g. development or production. LOG_LEVELstrGunicorn log level , e.g. debug, info, warning. WEB_TIMEOUTintGunicorn worker timeout in seconds. Workers silent longer than this are killed and restarted. WEB_WORKERSintNumber of Gunicorn worker processes . SECRET_KEYstrDjango secret key . Set a long random string in production. ALLOWED_HOSTSstrDjango allowed hosts . Use * for development; restrict to your domain in production. USE_X_FORWARDED_HOSTboolWhether to use the X-Forwarded-Host CACHE_BACKENDstrDjango cache backend , e.g. django.core.cache.backends.memcached.MemcachedCache. CACHE_LOCATIONstrDjango cache location , e.g. localhost:11211. SECURE_PROXY_SSL_HEADERstrDjango secure proxy SSL header . Provide as a comma-separated pair, e.g. HTTP_X_FORWARDED_PROTO,https.
Example defaults (from contrib/.env.sample): ENVIRONMENT =development LOG_LEVEL =debug WEB_TIMEOUT =60 WEB_WORKERS =2 SECRET_KEY =my-secret ALLOWED_HOSTS =* USE_X_FORWARDED_HOST =False CACHE_BACKEND =django.core.cache.backends.memcached.MemcachedCache CACHE_LOCATION =localhost:11211
Always replace SECRET_KEY with a strong, unique value before deploying to production.
Database Variable Type Description DATABASE_URLstrDatabase URL for the Django application. Must be PostgreSQL — Jarbas uses JSONField, which requires PostgreSQL.
Example: DATABASE_URL =postgres://jarbas:mysecretpassword@postgres/jarbas
When running with Docker Compose, also set POSTGRES_USER, POSTGRES_PASSWORD, and POSTGRES_DB to match your DATABASE_URL credentials: POSTGRES_USER =jarbas POSTGRES_PASSWORD =mysecretpassword POSTGRES_DB =jarbas
Message broker Variable Type Description CELERY_BROKER_URLstrCelery -compatible message broker URL. RabbitMQ is the default broker used by the project.
Example: CELERY_BROKER_URL =amqp://guest:guest@queue/
Used by the tweets management command to fetch Twitter engagement data for reimbursements.
Variable Type Description TWITTER_CONSUMER_KEYstrTwitter API key. TWITTER_CONSUMER_SECRETstrTwitter API secret. TWITTER_ACCESS_TOKENstrTwitter access token. TWITTER_ACCESS_SECRETstrTwitter access token secret.
Example: TWITTER_CONSUMER_KEY = TWITTER_CONSUMER_SECRET = TWITTER_ACCESS_TOKEN = TWITTER_ACCESS_SECRET =
Google settings Variable Type Description GOOGLE_ANALYTICSstrGoogle Analytics tracking ID, e.g. UA-123456-7. Leave empty to disable tracking. GOOGLE_STREET_VIEW_API_KEYstrGoogle Street View Image API key, used to display street-level photos for supplier locations.
Example: GOOGLE_ANALYTICS = GOOGLE_STREET_VIEW_API_KEY =
Production settings These variables are only required when deploying Jarbas with HTTPS termination via Let’s Encrypt.
Variable Type Description VIRTUAL_HOST_WEBstrHostname used for the HTTPS certificate, e.g. jarbas.serenata.ai. When testing production settings locally, add this hostname to /etc/hosts. LETSENCRYPT_EMAILstrEmail address used to register the HTTPS certificate with Let’s Encrypt. HTTPS_METHODstrSet to noredirect to disable the automatic HTTP → HTTPS redirect. Defaults to redirect.
Example: VIRTUAL_HOST_WEB =jarbas.serenata.ai LETSENCRYPT_EMAIL [email protected] HTTPS_METHOD =redirect
Searchvector schedule Jarbas can rebuild its full-text search index on a recurring schedule using Celery Beat. The schedule is defined with standard cron notation.
Variable Type Description SCHEDULE_SEARCHVECTORboolSet to True to enable scheduled searchvector rebuilds. SCHEDULE_SEARCHVECTOR_CRON_MINUTEstrCron minute field, e.g. 0. SCHEDULE_SEARCHVECTOR_CRON_HOURstrCron hour field, e.g. 2. SCHEDULE_SEARCHVECTOR_CRON_DAY_OF_WEEKstrCron day-of-week field, e.g. * for every day. SCHEDULE_SEARCHVECTOR_CRON_DAY_OF_MONTHstrCron day-of-month field, e.g. */2 for every other day. SCHEDULE_SEARCHVECTOR_CRON_MONTH_OF_YEARstrCron month-of-year field, e.g. * for every month.
Example (rebuild at 02:00 every other day): SCHEDULE_SEARCHVECTOR =True SCHEDULE_SEARCHVECTOR_CRON_MINUTE =0 SCHEDULE_SEARCHVECTOR_CRON_HOUR =2 SCHEDULE_SEARCHVECTOR_CRON_DAY_OF_WEEK =* SCHEDULE_SEARCHVECTOR_CRON_DAY_OF_MONTH =*/2 SCHEDULE_SEARCHVECTOR_CRON_MONTH_OF_YEAR =*
You can also trigger a one-off searchvector rebuild manually at any time:
docker-compose run --rm django python manage.py searchvector
New Relic (optional) Jarbas ships with New Relic APM integration. These variables are optional and only needed if you have a New Relic account.
Variable Type Description NEW_RELIC_APP_NAMEstrApplication name shown in the New Relic UI, e.g. Jarbas. NEW_RELIC_ENVIRONMENTstrDeployment environment label, e.g. Production. NEW_RELIC_LICENSE_KEYstrNew Relic account license key. NEW_RELIC_DEVELOPER_MODEstrSet to true to enable developer mode (no data sent to New Relic).
Example: NEW_RELIC_APP_NAME =Jarbas NEW_RELIC_ENVIRONMENT =development NEW_RELIC_LICENSE_KEY = NEW_RELIC_DEVELOPER_MODE =true
