Retry if provisioning target returns a 500

[MartyEwings]

Use Case

Frequently when provisioning nodes on GCP the service will return a 500, typically if the project is requesting many VMS at the same time.
This means if we provision 12 targets and one fails to provision, the entire workflow run fails, and we must examine if its a true failure or a provision failure. we would like it to retry on failure to mitigate this

Describe the Solution You Would Like

https://github.com/puppetlabs/provision/blob/main/tasks/provision_service.rb#L67-L82

This should retry a a fixed amount of times (2 failures in a row are uncommon) when it receives a 500 from the remote desitination

Describe Alternatives You've Considered

we have attempted looping the workflows in github actions on failures, but github actions really does not like this and our solutions are a little "hacky"