First and foremost, you clearly pick the wrong tools. The problem isn't "no-code/low-code". Tooks like Mendix, OutSytems, PowerApps, and more build on SQL Server, Oracle, and other decent datastores, use Java, .NET, etc., and are effectively robust code generators. I've built systems handling 300k users globally with low-code with zero issues that I would not expect to see in C# or Java (ie: a bad datamodel decision is always a bad datamodel decision and can impact performance).

Secondly, it seems like your team made terrible datamodeling choices and other mistakes as well. Did the tool make it easy for you to make those pool decisions? Sure. But the accountant who figures out a no-code tool isn't building an app to scale to that level of use, your leaders trusted you and your team to build a quality app, and 1) you chose the wrong tool 2) you make poor decisions, and your response is "no-code/low-code stinks".

I've built systems that were doing 70,000 credit checks *per hour* in low-code with zero issues. That sounds very scalable to me.

J.Ja