Write documentation on region failover decisioning

[devksingh4]

Should detail:

1. General failover model

• Active/Active except for Redis - theoretically we could run in active/active but S3 and other processes could have conflicts - easiest to just not do it

3. When to failover
4. Risks to failover:

• For example, in-progress locks are released if Redis is also down.
• S3 replication is not in time, so S3 data could be not accurately reflected in the new region in time.

4. How to failover (modify Cloudfront targets).

• provide an automated failover script

5. Prepare a disaster recovery account which can perform failover.