Bad scheduling checker

# Background
Sometimes pods can be scheduled in a non-HA way. For example, all pods in a deployment scheduled to a single machine, or a single zone. Authoring a deployment with HA is hard for beginners. We should be able to point out such risks and guide users to configure pod spread topology properly.

# Goals
* Detect bad scheduling issues
* All pods in a deployment scheduled to a single machine.
* All pods in a deployment scheduled to a single zone, if cluster supports AZ.
* Give general guidance about configuring pod anti-affinity and spread topology

# Non-goals
* No need to detect theoretical issues. Just check current status.

# References
* https://kubernetes.io/docs/concepts/scheduling-eviction/assign-pod-node/#affinity-and-anti-affinity
* https://kubernetes.io/docs/concepts/scheduling-eviction/assign-pod-node/#pod-topology-spread-constraints