Add retry logic for GetTargetResource()

[vhvb1989]

When looking for a service-host, azd searches a resource with the name of the service as a tag within the resource group.
If the Azure Resource Graph takes more time to update the index than the time it takes for azd to move from provision to deploy, azd would fail the deployment.

Azd should have a retry logic of at least 3 intents and an increasing delay time from 1, 3 and 5 seconds. Azd should then ask user if they want to keep trying to find the resource or give up, making the default option to give-up (so CI/CD is never blocked)

[weikanglim]

This makes sense! I've experienced similar transient errors in the past.

As a small mental note, we moved away from Azure Resource Graph to the actual ARM APIs awhile back, but I suspect sometimes tag/resource propagation may take awhile. I'm not sure which one might be the cause here.