Local and downstream cluster system component health is not validated in E2E tests

[nikkelma]

Changes made to simplify the cert-manager CRD installation process (#110) caused subtle issues with system-level components (#112) that were not detected in the existing E2E tests. Improve these tests to check known system components in the local cluster and the downstream quickstart-xx cluster(s) to confirm a fully healthy deployed quickstart setup.

Potential system components for health checks:

• cert-manager
• cattle agents (node, cluster)
• Ingress controller
• CoreDNS

Proposed test plan:

1. Update tests on master
2. Confirm tests pass for known healthy setup
3. Cherry-pick commit(s) onto 297bd4d
4. Confirm unhealthy cattle agent and cert-manager webhook causes test failure