DAQ Config Server

A service to put and get your config values from.

Comprises a FastAPI backend with a valkey database to C/R/U/D config values, and a Chakra/React frontend for easier management.

Currently the scope is JUST storing and retrieving feature flags for Hyperion/I03 UDC but we hope to expand this to replace all DAQ config files.

Source	https://github.com/DiamondLightSource/daq-config-server
Docker	docker run ghcr.io/DiamondLightSource/daq-config-server:latest
Releases	https://github.com/DiamondLightSource/daq-config-server/releases

A simple app for storing and fetching values. Has a Valkey (Redis) instance as well as options for file-backed legacy values (e.g. beamlineParameters...)

Currently the server application always needs to be run with the --dev flag, as it cannot yet look at the DLS filesystem to find the real beamline parameter files.

To use the config values in an experimental application (e.g. Hyperion) you can do:

from daq_config_server.client import ConfigServer

config_server = ConfigServer("<service ip address>", <port>)

use_stub_offsets: bool = config_server.best_effort_get_feature_flag("use_stub_offsets")

To work on the GUI you will probably need to run:

module load node
npm install

in the gui directory to setup the environment.

Testing and deployment

There is a convenient script in ./deployment/build_and_push_all.sh to build all the containers, which takes a --dev option to push containers with -dev appended to their names and a --no-push option for local development. This ensures that environment variables for dev or prod builds are included in the built container, such as the GUI pointing at the subdomain URL vs. localhost, and the root_path of the FastAPI app. To push to the registry you must have identified to gcloud by having loaded a kubernetes module and running gcloud auth login.

To deploy a live version, you can run the above script with no arguments and then while logged in to argus, in the daq-config-server namespace, run kubectl rollout restart deployment. If it is not currently deploy it you can deploy it with helm install daq-config ./helmchart.

To test locally, you can build everything with ./deployment/build_and_push_all.sh --dev --no-push and then run the containers daq-config-server-dev (with the command daq-config-server --dev), daq-config-server-db-dev, and daq-config-server-gui-dev, all with the --net host option.

To test on pollux, log in to pollux in your namespace and run:

helm install daq-config ./helmchart/ --values dev-values.yaml

followed by:

kubectl port-forward service/daq-config-server-svc 8080 8555

after which you should be able to access the frontend on http://localhost:8080 and the API on http://localhost:8555