A bit of backround: a service typically builds their API first, and then builds the console as a front-end to that API. For most services, everything you can do

A bit of backround: a service typically builds their API first, and then builds the console as a front-end to that API. For most services, everything you can do in the console, you can do via an API call. (I assume Honeycomb is an exception to this.) That API-first doctrine is at the heart of the underlying problems, which are: 1. Machine-friendly data models. 2. Distributed-friendly data models. 3. Deliberately independent and orthogonal services. 4. A vast range of customer needs over a console that doesn't itself make money. AWS really does put effort into making their console useful for getting work done. You'll find some workflows that are pretty usable. But when you get outside a handful of extremely common workflows that were technically convenient for a specific team to implement, those problems become dominant and the console becomes painful if not f*ing useless. That's why (I suspect) you hate it. > ... and staring at a spinner spin for far too long. Oh yeah, fundamental cloud problem #5: "A few minutes" is absolutely forever when you're troubleshooting.