Configuration

Most settings can be configured from the frontend. A few must be set in the .env file.

For interactive setup:

bash gen-env.sh

Or manually copy .env-sample to .env and edit.

Secure Cookies

Bayanat uses secure cookies by default (requires HTTPS). For development:

SECURE_COOKIES=False

DANGER

Do not disable secure cookies in production.

Secret Key

SECRET_KEY keeps sessions secure. Generate a strong key:

python3 -c "import secrets; print(secrets.token_hex(32))"

WARNING

Changing the secret key logs out all users.

PostgreSQL

Required for Docker deployment. Optional for native installs on the same host.

• POSTGRES_DB: Database name (default: bayanat)
• POSTGRES_HOST: Host (empty for local, postgres for Docker)
• POSTGRES_PASSWORD: Password (not required for local)
• POSTGRES_USER: Username (not required for local)

Redis

Required for Docker deployment.

• REDIS_HOST: Host (empty for local, redis for Docker)
• REDIS_PASSWORD: Password, if set

Password Salt

SECURITY_PASSWORD_SALT must be generated and kept secret:

python3 -c "import secrets; print(secrets.token_hex(32))"

DANGER

Changing the password salt invalidates all user passwords.

Two-Factor Authentication

Generate a TOTP secret for SECURITY_TOTP_SECRETS:

python3 -c "import secrets; print(secrets.token_hex(32))"

See Flask Security docs.

DANGER

Changing this secret invalidates all 2FA configurations.

Storage

Local

Media files stored in enferno/media/ relative to the application directory. An installer-managed v5 host keeps them in /opt/bayanat/shared/media, which the release symlinks into place, so the application path is the same either way.

Amazon S3

Configure the S3 bucket with correct policies, block public access, and set up CORS.

S3 does not hold everything

Uploads made through the interface go straight to the bucket, but inline images in descriptions do not. They are always written to enferno/media/inline/ on the server, with no S3 path, so bucket versioning and replication do not cover them.

Back up that directory as well as the bucket. Imported files are staged locally and removed once uploaded, so they need no separate handling.

Search

Interactive searches run under a database statement timeout. When a search exceeds it, the query is cancelled and re-run by a background worker instead of failing, and the user is notified when the results are ready. See Search for what this looks like in the interface.

Variable	Default	Purpose
SEARCH_TIMEOUT	30	Seconds an interactive search may run before it is handed to the background. 0 disables the behaviour and searches run unbounded.
BACKGROUND_SEARCH_TIME_LIMIT	600	Seconds the background re-run may take before it is abandoned.

Background searches require a running Celery worker. Without one, users receive the "continuing in the background" message but never get results.

WARNING

Lowering SEARCH_TIMEOUT far below the default sends ordinary searches to the background, including the initial page load of a list view. Raise it instead if legitimate searches are being deferred.

Sessions and Login Throttling

Variable	Default	Purpose
SESSION_LIFETIME	3600	Seconds of inactivity before a session expires and the user must sign in again.
LOGIN_RATE_LIMIT_PER_USERNAME	10 per 15 minutes	Failed-login throttle applied per account, so one targeted account cannot be brute-forced from many addresses.
LOGIN_RATE_LIMIT_PER_IP	30 per 15 minutes	Failed-login throttle applied per source address, so one address cannot spray many accounts.

Both throttles use the limits string syntax, for example 5 per minute or 100 per hour. They apply to the login endpoint only. Lowering SESSION_LIFETIME too far is a common cause of complaints about constant re-authentication.

Configuration File Location

Variable	Default	Purpose
BAYANAT_CONFIG_FILE	config.json	Path to the feature-toggle configuration file.

Installer-managed deployments set this to a path outside the release directory, because releases are read-only to the services and config.json is written at runtime from the admin interface. Set it explicitly if you deploy releases as read-only trees.

Data Import

Enable path scanning with ETL_ALLOWED_PATH.

WARNING

Only enable path scanning when needed.

Backups

See Backups for configuration.

Offline Maps

For offline or privacy-focused deployments, run your own tile server using openstreetmap-tile-server. Update the Maps API Endpoint in system settings.