4 ms·
If companies only listened to the CTO of the infra supplier they use: 'Everything breaks all the time, design your infra with that in mind' AWS has multiple re
by mvanbaak 6y ago
If companies only listened to the CTO of the infra supplier they use: 'Everything breaks all the time, design your infra with that in mind'
AWS has multiple regions, and within a region multiple availability zones.
If you are as big as roomba, you should simply have your api's running in more than one region. Then, if us-east-1 goes down (which happens from time to time) your users will not notice (ok, the geeks will because latencies will change)
So yeah, while the underlying problem was an outage of a region of AWS, it's not AWS that failed in this case. People should simply design their applications like AWS is telling them.