Spike-Research Federation Failover for TP

Gather all the use cases and try to prioritize the most common or most urgent ones to document for Tech Preview

QE Test Case doc -

https://docs.google.com/document/d/1P-34apfqEUeyOwjDUcKt63oxVB9c0_2AIwr1rrAUPS8/edit#heading=h.bki0pcsuu9qt

 Simple cases

• Local pod failures - do they fail over to remote service 
  • When a service fails in one mesh does the traffic get routed to the same service running in another service mesh peer? If so, what does the customer have to configure to make this happen?
• Local node failures - does traffic for workloads running on downed node get routed to other mesh
  • When a node in one mesh fails does the traffic get routed to similar services running in another service mesh peer? If so, what does the customer have to configure to make this happe

Complicated cases (not for 2.1. Broke story and captured complicated cases in OSSMDOC-401)

• API server is degraded and istiod can’t derive local workloads for services

• Mesh Failure (Service failure in the Mesh - Ex: kill Ratings pod)