An availability zone (AZ) in AWS eu-west-2 was flooded by a fire protection system going off within the last year. It absolutely did affect workloads in that AZ. That shouldn't have had a large impact on their customers since AWS promote and make as trivial as is viable multi-AZ architectures.
Put another way: one is guided towards making operational good choices rather than being left to discover them yourself. This is a value proposition of public clouds since it commoditises that specialist knowledge.
What surprised me most about today's fire is that their datacenters have so little physical separation. I expected them to be far enough apart to act as separate availability zones.
Put another way: one is guided towards making operational good choices rather than being left to discover them yourself. This is a value proposition of public clouds since it commoditises that specialist knowledge.