4 ms·

That will increase the chances of cross-region failures though.

• robryan 15y ago
  Almost need an availability zone built for a failure event that otherwise sees little use, or just spot use. So then you can assume that the people on it need the same resources their live implementation is taking up before failure so there is no overload when the failure occurs.
  • justincormack 15y ago
    That would mean a 25% price hike on all services. Rather a lot for a situation when the biggest issues seem to be lack of preparedness and planning for failure by a lot of people.