Use grpc time outs and ability to wait on jobs to have the server hang pending
on a job coming through. This reduces the need for constant update requests
while still responding to new jobs quickly.

This is part of #118 and #126

Bot will periodically attempt to reconnect if a controller goes down.
If a job is running it will similar to the server side, wait on the job
completing else respond periodically.

This logic may need changing if/when bots can have multiple leases.

This is part of #118 and #126

This stops the issue of grpc calls timeouts defaulting to the max int64 value.
MAX_JOB_BLOCK_TIME is in the settings file.