Was that competitor priced competitively with AWS? I think of the project management triangle here - good, fast, or cheap - pick two. AWS would be fast and cheap.
Yes, good point. Pricing is a bit higher. As another reply pointed out: there's ~three that work on the same scale. This was one, another hint I guess: it's mostly B2B. Normal people don't typically go there.
Azure, from my experience with it has stuff go down a lot and degrades even more. Seems to either not admit the degradation happened or rely on 1000 pages of fine print SLA docs to prove you don't get any credits for it. I suppose that isn't the same as "lose a region resiliency" so it could still be them given the poster said it is B2B focused and Azure is subject to a lot of exercises like this from it's huge enterprise customers. FWIW I worked as a IaC / devops engineer with the largest tenant in one of the non-public Azure clouds.
My $3/mo AWS instance is far cheaper than any DIY solution I could come up with, especially when I have to buy the hardware and supply the power/network/storage/physical space. Not to mention it's not worth my time to DIY something like that in the first place.
False equivalence/moving goalposts IMO... I was only refuting your claim of "AWS is not cheap", as if it's somehow impossible for it to be cheap... which I'm saying isn't the case.
Sorry to jump in y'alls convo :) AWS is cheaper than the Cloud we built... I just don't think it's significant. Ours cost more because businesses/governments would pay it, not because it was optimal.
Price is beside my original point: Amazon has enjoyed decades for arbitrage. This sounds more accusatory than intended: the 'us-south-1' problem exists because it's allowed/chosen. Created in 2006!
Now, to retract that a bit: I could see technical debt/culture making this state of affairs practical, if not inevitable. Correct? No, if I was Papa Bezos I'd be incredibly upset my Supercomputer is so hamstrung. I think even the warehouses were impacted!
The real differentiator was policy/procedure. Nobody was allowed to create a service or integration with this kind of blast area. Design principles, to say the least. Fault zones and availability zones exist for a reason beyond capacity, after all.
