Add logs to better understand slow results retrieval

[kt474]

Debugging slow results retrieval has been difficult across the whole stack. We don't believe the issue is on the api side so we should add debug logs on the client side to help pinpoint the issue.

• Request/reply debug-level logs on the client-side (specifically: response time)
• Consider logging serialization time (each major step in job.result() needs to be logged)
• Since this issue is difficult to replicate this will give users a way forward when they encounter issue again

[kt474]

Closing this for now as we are seeing much better job results retrieval after #653