Use kubectl wait to make the integration test more reliable

[lisguo]

Issue #, if available: N/A

Description of changes:

• Instead of using sleep, we should use kubectl wait for resources to be available before continuing with the test.
• Delete pods instead of using kubectl rollout which will stagger pods, the wait command to finish early as pods will be in Ready state before restarting all pods.
• Rename AmazonCloudWatchAgentIntegrationTest to InstrumentationTest

This error message is common in the test case for AmazonCloudWatchAgent agent pod creation step:

Error from server (InternalError): error when creating "integration-tests/manifests/cloudwatch-agent-daemonset.yaml": Internal error occurred: failed calling webhook "mamazoncloudwatchagent.kb.io": failed to call webhook: Post "[https://cloudwatch-webhook-service.amazon-cloudwatch.svc:443/mutate-cloudwatch-aws-amazon-com-v1alpha1-amazoncloudwatchagent?timeout=10s](https://cloudwatch-webhook-service.amazon-cloudwatch.svc/mutate-cloudwatch-aws-amazon-com-v1alpha1-amazoncloudwatchagent?timeout=10s)": dial tcp 10.110.30.107:443: connect: connection refused

Example run: https://github.com/aws/amazon-cloudwatch-agent-operator/actions/runs/9552268986/job/26328442298?pr=186

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

[lisguo]

this is the main reason why it was flaky. operator wasn't ready to serve on the webhook.