We have decided that having a daemon server for tierkreis is desirable as such we should start writing a draft API specification for the endpoints for a server that the user can interact with. Ideally this server will allow users to interact with the daemon and provide enough information for the existing visualization or debugging.
A non-exhaustive list of capabilities:
- Upload a workflow to the daemon
- Request an uploaded workflow
- Start the workflow with a set of parameters, creating a "workflow run"
- Retrieve the state of a workflow run
- Retrieve the output of a workflow run
Many of these capabilities will exist in the visualization server already, but it would be useful to decide on a specification to reduce duplication.
We have decided that having a daemon server for tierkreis is desirable as such we should start writing a draft API specification for the endpoints for a server that the user can interact with. Ideally this server will allow users to interact with the daemon and provide enough information for the existing visualization or debugging.
A non-exhaustive list of capabilities:
Many of these capabilities will exist in the visualization server already, but it would be useful to decide on a specification to reduce duplication.