Don't Permanently Fail Node Thread on Unreachable AMI during Start

[jxmx]

Describe the request
When starting the allmon3 process, if the AMI process is not contactable, don't fail the worker thread for that node but instead retry according to the retry logic.

There was I reason I wasn't doing this before and I need to figure out why and if that's still valid.