Scale Fail (part 2)

everything has downtime. there are no 100% solutions. your bank's site will be down. the stock market suspends trading. power to your house fails. no water comes out of the faucet.

remember amazon's last big outage? what was the date? bet you can't tell me unless you look it up. the public moves on, there isn't much to be gained by engineering a path around these scenarios.

engineering around certain types of failure states is pointless, you create a huge opportunity cost with regard to allocating resources to new features that make your service more attractive. being pathological about availability is unrealistic and can be deadly to a business.