Configuration files#
Copy .env.example to .env in the twin repository. Set GIS_REGION_CONFIG to a region YAML file. Docker Compose reads .env; the website does not read it.
Never commit deployment secrets. Use .env.example as the inventory of published environment settings and inspect the module source for additional model settings and defaults.
Main settings#
| Setting | Purpose |
|---|---|
GIS_REGION_CONFIG | GIS catalogue, layers and extent |
MARKET_DISPATCH_URL, MARKET_REGION | Market adapter endpoint and region identifier |
HYDRO_URL, HYDRO_SITES | Hydrometry endpoint and selected gauges |
STORAGE_URL, STORAGE_BBOX | Storage endpoint and local extent filter |
CONSOLE_PORT, CONSOLE_BASEMAP | Local console port and optional starting basemap |
SIM_MAX_PIPES, SIM_MAX_SNAP_M | Supplied subgraph cap and reservoir snap limit |
SIM_RESIDUAL_MIN_SAMPLES | Minimum residual baseline count |
SIM_RESULT_RETENTION_HOURS | Per-element result retention |
Changing the model subgraph settings requires applying them to both the seeder and simulator, then reseeding and rebuilding. Reseeding replaces synthetic registries.
Console secrets#
./scripts/init-secrets.shAdd --totp when you want an authenticator secret. The script fills blank entries without printing the secrets. Public admin access requires the password hash, relay URL and separate admin token. Set secure cookies for an HTTPS deployment and configure allowed hosts appropriately.
External Elasticsearch#
Use ELASTICSEARCH_URL, network, credentials and CA settings for an existing cluster. Run scripts/elk-up.sh to select and validate the target. A TLS or authentication error needs correction; do not bypass the script by starting Logstash directly.
Literal Flink settings#
Flink SQL does not substitute environment variables. Its JDBC URL and credentials in flink/sql/02_sinks.sql must agree with Compose. Logstash does support environment substitution. Changing .env alone does not rewrite Flink SQL.