[meta] Create outage response plan for v3

## Summary

Create a plan for when things break. For example:

* How will we identify an incident?
* Can we dogfood Elastic Synthetics?
* Roles and responsibilities?
* Comms plan?
* On-call plan?
* Use pager duty?
* RCA plan?
* Resolution paths?
* Post-incident actions?
* Sev levels?
* Testing?
* If we're down, how do we get a page up?

This plan should cover as much of https://elasticco.atlassian.net/browse/PRR-4931 as possible. Some requirements in the PRR will be out of scope by design.

## Scope

While this is needed for V3, our legacy docs, and our API docs, V3 will be the primary initial scope. The outage response plan should cover basic steps for resolving issues with bump (contacts, etc.) but should be mainly focussed on V3. Legacy docs are entirely out of scope for now.