Batch pool can resize continuously if nodes never become usable

[MattMcL4475]

Describe the bug
If the Batch pool never is able to successfully scale up, it may continuously attempt to do so.

Steps to Reproduce
Create a Network Security Group for the Batch node subnet that has all outbound traffic blocked. This results in the Azure Batch service's node agent not able to call home, which results in the node becoming unusable

Expected behavior
TES should stop attempting to scale the pool back up, and TBD potentially stop scheduling tasks and fail Create Task API requests with a 503 error and include in the response a sentence such as "Permanent Azure Batch scaling issue - please contact your system administrator."