Most settings can be configured by creating a custom Docker environment file src/docker/.env (see src/envs/*.env for options). You can also edit docker-compose.yml and /etc/docker/daemon.json, but we recommend sticking to .env.
Integration settings that deal with external systems such as TLS, accounts, and notebooks require having launched them. If you are not integrating into existing ones, see the initial setup section for how to quicklaunch a Graphistry server.
By default, graph-app-kit logging uses the Docker json file driver:
-
Inspect recent activities:
cd src/dockerand thensudo docker-compose logs -f -t --tail=100
- Streamlit: URL base path:
BASE_PATH=dashboard/andBASE_URL=http://localhost/dashboard/ - Graphistry: None - set
GRAPHISTRY_USERNAME=usr+GRAPHISTRY_PASSWORD=pwd(seesrc/envs/graphistry.envfor more, likeGRAPHISTRY_SERVERif using a private Graphistry server) - Log level:
LOG_LEVEL=ERROR(for Python'slogging)
- Amazon Neptune guide for TinkerPop/Gremlin integration
- Auth: See Caddy sample reverse proxy example for an authentication check against an account system, including the one shipping with your Graphistry server (requires
sudo docker-compose restart caddyin your Graphistry server upon editing/var/graphistry/data/config/Caddyfile)
- To simulatenously run 1 public and 1 private instance, create two
graph-app-kitclonespublic_dashandprivate_dash, and forsrc/docker/.env, set:COMPOSE_PROJECT_NAME=streamlit-pubandCOMPOSE_PROJECT_NAME=streamlit-priv- Override default
ST_PUBLIC_PORT=8501with two distinct values
- See Caddy sample for configuring URI routes, including covering the private instance with your Graphistry account system (JWT auth URL)