When running on a FIPS-enabled cluster, the e2e test TestFirstBootHasSSHKeys times out.

Always

1. Open a PR to the MCO repository.
2. Run the e2e-aws-ovn-fips-op job by commenting /test e2e-aws-ovn-fips-op (this job does not run automatically).
3. Eventually, the test will fail.

=== RUN TestFirstBootHasSSHKeys1065mcd_test.go:1019: did not get new node
--- FAIL: TestFirstBootHasSSHKeys (1201.83s)    

=== RUN   TestFirstBootHasSSHKeys
    mcd_test.go:929: Got ssh key file data: XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
        
--- PASS: TestFirstBootHasSSHKeys (334.86s)

It looks like we're hitting a 20-minute timeout during the test. By comparison, the passing case seems to execute in approximately 5.5 minutes.

I have two preliminary hypothesis' for this:
1. This operation takes longer in FIPS-enabled clusters for some reason.
2. It is possible that this is occurring due to a difference in which cloud these tests run. Our normal e2e-gcp-op tests run in GCP whereas this test suite runs in AWS. The underlying operations performed by the Machine API may just take longer in AWS than they do in GCP. If that is the case, this bug can be resolved as-is.