After the outage of EC2 in April, Netflix was one of the few sites that was still running and attributed it to the architecture that was being considered to be resistant to failure (https://medium.com/netflix-techblog … ). This considered, how is it that Netflix was affected by the much less severe disruption of EC2 on August 8th,and not the April 21 EC2 outage?
Hi Judd,
It is indeed true that Netflix withstand the EC2 outage of April 21 but was taken down on August 8 EC2 outage. The reason for that is because; April 21 EC2 outage only affected a single and specific availability time zone in the US-EAST 1 region. Whilst the EC2 outage on August 8 affected the entire US-East 1 region and all availability time zones. Their architecture is designed to withstand whole zone outage theoretically in becoming unavailable. But they have not designed the system and architecture to withstand all and whole regions going offline. Also added to this fact is that there are also some issues on Amazon Relational Database Service in Virginia.
Aristono
