Hacker News new | ask | show | jobs
by rozenmd 1428 days ago
There were zero issues with performance/reliability that caused the outage, they just deleted hundreds of customers data, and had to manually restore the data due to lack of automated restore scripts.
1 comments

Reliability is measured from a customer's perspective, not from a geeky vendor-oriented perspective of a backend service. When the vendor deletes data, the vendor's reliability quotient decreases, because that data is no longer available to the customer.
Once again we learn the hardware that accessibility is not the same thing as reliability.