speed up test runs

[mattfenwick]

client-side:

• https://github.com/kubernetes/kubernetes/blob/master/test/images/agnhost/connect/connect.go

server-side:

• https://github.com/kubernetes/kubernetes/tree/master/test/images/agnhost/porter
• https://github.com/kubernetes/kubernetes/tree/master/test/images/agnhost/serve-hostname

ideas:

• reduce number of kubectl exec calls
• avoid spinning up an agnhost binary for each network call

[lwr20]

A couple more possible ideas (though note I've not read the code in this area, so bear that in mind):

• run checks in parallel
• run all the checks continuously within the pod and report current result via an http server (or equivalently a prom metrics endpoint) - so checking the overall connectivity map becomes a single curl to each pod
• run a pod for each check that continuously does the check and report success/failure via its Readiness check? Then checking the overall connectivity map is simply a kubectl get po -A (or equivalent api call)