add fingerprint retry configuration knobs

[tgross]

When a Nomad client starts, it fingerprints its environment. In cloud environments (AWS, GCP, Azure, Digital Ocean), this includes hitting instance metadata endpoints. But in sufficiently fast instance startup, it's possible that the cloud provider's instance metadata endpoint simply isn't available by the time the Nomad agent starts. We currently timeout these requests fairly quickly to detect that we're not in that fingerprinter's cloud environment. But most Nomad agent configs will be crafted for their specific environment and the cluster admin will add other clouds to the denylist.

In those circumstances, it would make sense to allow the fingerprint to retry for a much longer configurable time and potentially even fail the agent startup if it means the client won't be able to properly run.