At the moment the only way to pass env vars to a running instance of Llama Deploy is through setting them before starting the process, which is not practical in case of a remote installation or docker containers.
We should add the option to declare the environment variables needed by each service like this:
At the moment the only way to pass env vars to a running instance of Llama Deploy is through setting them before starting the process, which is not practical in case of a remote installation or docker containers.
We should add the option to declare the environment variables needed by each service like this:
then the Control plane would propagate the env var when spinning up the service.