Device manager improvements

[MagdaPaj]

• Setup the Device Manager to run as a generic host (there is many things provided out-of-the-box, so the setup is simpler)
• Configure EventHubReceiverService as IHostedService so StartUp and Shutdown are handled by the Host.
• Some small updates to EventHubReceiverService for fixing warnings or improving readability.

As a result of that:

• we don't see any more the Device Manager pod being restarted every 5 min
• there is still a weird behavior in the outgoing messages metric reported by Event Hub after stopping the Device Manager and starting it, but it comes back to normal
  
• there are still some errors reported from the BlobStorage client (404) but it happens less frequently as the application is not restarted every 5 min
  
• I also validated checkpointing and it behaves correctly, it starts where it previously finished. And the number of calls to the checkpoint container matches the number of requests from the Device Manager to the Device API (though there is a bug and all telemetry reporting HTTP requests are duplicated, but if we write a query correctly then we can see it matches
  

• I found out that the huge amount of dependencies calls comes from the EventProcessorClient that uses Azure Storage blob to manage ownership. There is one PUT and one GET request every 30 seconds to the ownership Blob container, even when we're not processing any messages. Originally I thought that we were making so many calls to checkpoint Blob container, but that's not the case.
  

This doesn't solve but decreases the number of errors from this ticket.

[MahrRah]

LGTM