Well, that's a pretty self defeating argument. As I've said, you cannot scale on http requests (which makes me think you don't know the difference between websockets vs http)... unless you want to make it more slow and eventually dos it. If you're so familiar with scaling challenges, you'd have agreed with me that web sockets is the way to go... and it's not that this is not known and being used by others. Renting more servers (scaling) does nothing when your fundamentals (networking stack) are not right. You cannot build a house on sand, build new rooms next to it to increase capacity and expect it will get better than it get worse. Just my 2 cents.
in the end you are saying that Kraken is too incompetent on running a trade platform. If it is so easy, then the problems should have been solved long ago, as these problems started months of not years ago. Even when they would have had to get a new dev team (which I strongly advise to Kraken, as the current one then seems too incompetent when drawing conclusions from your statement) they should have had sufficient time to scale there systems according to trading volumes etc