INSTALL, conformance-test scripts: apply configuration for aws-k8s-cni v1.6.0

[etungsten]

Issue #, if available: N/A

Description of changes:
Add install instructions to apply configuration for v1.6.0 of the aws-k8s-cni plugin
Updated sonobuoy conformance test script to apply v1.6.0 of the aws-k8s-cni plugin to the test cluster

Testing done:
Applied configuration for aws-k8s-cni v1.6.0 on a fresh cluster and ran sonobuoy conformance tests over it, tests pass:

✦ $ sonobuoy results 202002140125_sonobuoy_2ed81549-a7da-442a-99f2-5b6f1973a3d3.tar.gz
Plugin: e2e                       
Status: passed                                                                       
Total: 3585                                                                                   
Passed: 204                                                                                                                                    
Failed: 0                                                                             
Skipped: 3381

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

[tjkirch]

It does make sense to me to apply the full config relevant to the version instead of just updating the URI, because they've changed a couple other settings that could be important. I'm most worried about the addition of the health probe; it could be a good thing, but we haven't run with it before. @samuelkarp - do you have any experience with this?

[samuelkarp]

@tjkirch What worries you specifically about the probes?

In general, probes exist to provide information about the state of the application running inside a container; a readiness probe is used to determine whether the application has started and is ready to perform work while a liveness probe is used to determine whether a container should be restarted.

You can find more about probes here and how they interact with the pod lifecycle here.

In this case, I don't see a restartPolicy specified, so I don't know that the liveness probe is going to have any effect. However, I would bias toward assuming that the settings defined there are fine, in that they're specified by the EKS team for use with EKS and other Kubernetes clusters running on AWS.

[tjkirch]

@tjkirch What worries you specifically about the probes?

@samuelkarp Nothing worries me specifically, I just had no experience with them in production. I know what they're for, but don't know if they're likely to cause us problems. I'll assume you think not, given your comments and approval.

[samuelkarp]

I'll assume you think not, given your comments and approval.

Discussed offline yesterday, but I failed to follow up. Yes, I think this is fine.