Authy’s Fault-Tolerant Infrastructure
Last week, Lean Stack did a published a great interview where we talked in-depth how we built our infrastructure. Our primary goal when designing our infrastructure was uptime. Other big considerations were security and speed.
Uptime is usually expressed as a percentage of availability of your application per year. Our uptime goal was 99.999% availability each year. This roughly means we could have a max downtime of 31.5 seconds for the full year. Obviously, accomplishing this meant we had to automate failover completely. That’s exactly what we did and so far we’ve been able to meet this uptime.
You can read more about it here.