Create SRE Guide for Kubernetes

[christopherhein]

Add a module discussing tactics around: debugging, outages, failures in distributed systems, etc.

[arun-gupta]

EC2 instance type: https://kubernetes.io/docs/admin/cluster-large/

[StevenACoffman]

Related: How to Recover a Broken Kubernetes Cluster

[christopherhein]

https://kubernetes.io/docs/tasks/configure-pod-container/configure-service-account/#use-the-default-service-account-to-access-the-api-server

[StevenACoffman]

Also, avoid t2 (burstable) instance types, as Kubernetes overschedules them when they have spare capacity and then they become unresponsively slow when their burst credit is used up. This leads to cascading failures, as things get re-scheduled to responsive nodes that then use up their credits.