fix(nodeadm): increase IMDS retries

[cartermckinnon]

Description of changes:

This increases the number of retry attempts for IMDS requests from 15 to 60. We use exponential backoff with a max delay of 1 second, so this results in about 1 minute of retries. We've seen rare failures where the network is not up until the previous retries (15) were exhausted.

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

[ndbaker1]

thanks!

[ndbaker1]

any value in exercising a longer downtime in the e2e test?

amazon-eks-ami/nodeadm/test/e2e/cases/imds-timeouts/run.sh

Lines 12 to 23 in b5c6c83

	# configure without launching the imds mock service
	IMDS_MOCK_ONLY_CONFIGURE=true mock::aws
	if nodeadm init --skip run; then
	echo "bootstrap should not succeed when EC2 IMDS APIs are not reachable."
	exit 1
	fi
	# start the imds mock part way into the initialization to mimic
	# delayed availability of IMDS
	{ sleep 10 && AWS_MOCK_ONLY_CONFIGURE=true mock::aws; } &
	nodeadm init --skip run