Continued wrenching on the serverless relay this week:
- Removed Cloudfront distributions from the equation, now using a custom domain with Route 53 latency based routing policies
- RDS Postgres replica is consistently < 1 second replication lag upon db write between the primary instance (us-east-1) and replica (eu-central-1)
- Synthetic tests run from US and EU are running every minute and producing pretty consistently results (still need to try to cut down latency from lambda invocations)

- Still getting getting a fair number of 410 GoneException errors when replying to clients, continuing to dig into this :(