Tink Server
Tink Server exposes workflow actions over a gRPC API for Tink Worker to retrieve and execute on a machine being provisioned. When a Tink worker completes an action, it reports the status back to the Tink Server and retrieves the next action, if any. The Tink Server uses the Workflow Kubernetes resource, continuously updating the Status based on the Worker’s reports.
The Tink Server <-> Worker communication protocol is implemented with Protobuf. The proto files can be found here.