Configuration
The web app is configured with environment variables. Copy ex-accell-ui/.env.example to .env to get started.
Database
Use either DATABASE_URL or the individual settings. DATABASE_URL takes effect when set.
| Variable | Default | Description |
|---|---|---|
DATABASE_URL | none | Full connection string, e.g. postgresql://user:pass@host:5432/analytics |
POSTGRES_HOST | localhost | Database host |
POSTGRES_PORT | 5432 | Database port |
POSTGRES_DB | none | Database name |
POSTGRES_USER | none | Database user |
POSTGRES_PASSWORD | none | Database password |
POSTGRES_POOL_MAX | pool default | Maximum pooled connections |
Authentication
| Variable | Default | Description |
|---|---|---|
SESSION_SECRET | generated in dev | Signing key for session cookies, at least 32 characters. Required in production. Generate with openssl rand -hex 32. In development an ephemeral key is generated, so sessions end when the server restarts. |
SESSION_TTL_HOURS | 12 | How long a sign-in lasts |
COOKIE_SECURE | true in production | Set to false only when serving plain HTTP outside localhost |
ADMIN_EMAIL | none | Email of the first admin, created when no users exist |
ADMIN_PASSWORD | none | Password for that admin, at least 12 characters |
ADMIN_NAME | Administrator | Display name for that admin |
Server
| Variable | Description |
|---|---|
PORT | Port for the production server (node build) |
ORIGIN | Public URL of the app (for example https://analytics.example.com). SvelteKit uses it to build URLs and for its cross-site request check, so set it whenever the server sits behind a proxy |
BODY_SIZE_LIMIT | Largest request body the server accepts, in bytes (suffixes K, M, G work). Defaults to 512K; larger requests get a 413 before any route runs. Raise it only if you need bigger workflow definitions or uploads |
API_RATE_LIMIT_PER_MINUTE | How many /api/* requests one signed-in person may make a minute. Defaults to 1200; 0 turns the limit off. Over the limit the server answers 429 with a Retry-After header. The count is kept in memory, per server process |
LOG_LEVEL | debug, info (default), warn or error |
NODE_ENV | development or production |
Data sharing (IPFS)
Off by default. Read Data sharing over IPFS before turning it on: anything published is public on the IPFS network; an administrator can stop your node serving it, but copies others fetched cannot be recalled.
| Variable | Default | Description |
|---|---|---|
IPFS_ENABLED | false | Must be exactly true to turn peer-to-peer sharing on; any other value leaves it off |
IPFS_BLOCKSTORE_PATH | /data/ipfs/blocks | Where the node keeps published and fetched blocks |
IPFS_DATASTORE_PATH | /data/ipfs/datastore | The node's identity and routing data; keep it between restarts |
IPFS_GATEWAY_URL | https://ipfs.io/ipfs | Public gateway used to build shareable links; the server never fetches from this address (the node itself does use other public gateways, see the guide) |
IPFS_FETCH_TIMEOUT_MS | 120000 | How long an import waits for the network before answering 504 |
IPFS_MAX_FETCH_BYTES | 268435456 | Largest package an import accepts (256 MB); bigger ones get 422 |
PACKAGES_DIR | /data/packages | Where generated and imported data packages are written |
IPFS_LISTEN_ADDRS | libp2p defaults | Comma-separated multiaddrs the node listens on, replacing the default random ports (for example /ip4/0.0.0.0/tcp/4001) |
IPFS_ANNOUNCE_ADDRS | none | Comma-separated multiaddrs announced to peers instead of the detected ones |
IPFS_UPNP | true | false stops the node asking the router to open ports |
IPFS_RELAY_SERVER | true | false stops the node relaying other peers' traffic |
IPFS_AUTOTLS | true | false stops the node requesting a libp2p.direct TLS certificate |
MAX_EXPORT_ROWS | 500000 | Most rows exported in one publish and loaded in one import |
Keep secrets out of git
.env files are ignored by git. Never commit real credentials. CI scans every push for secrets.