I don't really have much to do with contracts - but my company is stating that we have up time of 99.xx%.
In terms of contract customers don't care if I have Azure/AWS or I keep my server in the box under the stairs. Yes they do due diligence and would not buy my services if I keep it in shoe box.
But then if they loose business they come to me .. I can go after Azure/AWS but I am so small they will throw some free credits and me and tell to go off.
Maybe if you are in B2C area then yeah - your customers will probably shrug and say it was M$ or Amazon if you write sad blog post with excuses.
It's going to depend on the penalties for being unavailable. Small B2B customers are very different from enterprise B2B customers too, so you ultimately have to build for your context.
If you have to give service credits to customers then with "one box" you have to give 100% of customers a credit. If your services are partitioned across two "shards" then one of those shards can go down, but your credits are only paid out at 50%.
Getting to this place doesn't prevent a 100% outage and it imposes complexity. This kind of design can be planned for enterprise B2B apps when the team are experienced with enterprise clients. Many B2B SaaS are tech folk with zero enterprise experience, so they have no idea of relatively simple things that can be done to enable a shift to this architecture.
Enterprise customers do care where things are hosted. They very likely have some users in the EU, or other locations, which care more about data protection and sovereignty than the average US organization. Since they are used to hosting on-prem and doing their own due diligence they will often have preferences over hosting. In industries like healthcare, you can find out what the hosting preferences are, as well as understand how the public clouds are addressing them. While not viewed as applicable by many on HN due to the focus on B2C and smaller B2B here, this is the kind of thing that can put a worse product ahead in the enterprise scenario.
It really varies a lot. I have seen very large lazy sites suddenly pick up a client that wanted RCA for each bad transaction, and suddenly get religion quickly (well quickly as a large org can). Those are precious clients because they force investment into useful directions of availability instead of just new features.
Because you have a vendor/customer relationship. The big thing for AWS is employer/employee relationships. If you were a larger company, and AWS goes down, who blames you? Who blames anyone in the company? At the C-level, does the CEO expect more uptime than Amazon? Of course not. And so it goes.
Whereas if you do something other than the industry standard of AWS (or Azure/GCP) and it goes down, clearly it's your fault.
I don't really have much to do with contracts - but my company is stating that we have up time of 99.xx%.
In terms of contract customers don't care if I have Azure/AWS or I keep my server in the box under the stairs. Yes they do due diligence and would not buy my services if I keep it in shoe box.
But then if they loose business they come to me .. I can go after Azure/AWS but I am so small they will throw some free credits and me and tell to go off.
Maybe if you are in B2C area then yeah - your customers will probably shrug and say it was M$ or Amazon if you write sad blog post with excuses.