This currently contains the logic for restarting updated jobs,
but nothing exercises this logic. The logic for starting polling
for a newly created pipeline is implemented.
This includes refactoring the jobscheduler into the runnermanager. This
service manages runner connections and allocating them to pipelines.
These requests are done via the pipeline grpc api
This container's network is configured based on parameters in the config
file. If configured correctly, this will allow the pipeline script to
speak to the cursorius server over the pipeline api.
Server now accepts runner connections via websockets. Jobs
will be delegated to these runners. This is currently a WIP,
as it is still hardcoded to use the first available runner.