[zedpm]

The AWS issue in US-West-2 is only affecting a single AZ, according to them. A single AZ should never take down a production service.

[codeduck]

AWS single az failures can frequently cause general failures across other AWS services like lambda, Kinesis etc. Your app might tolerate the failure but be taken out by events beyond your control.

Ideally you'd want a way to fail traffic over to another region.