Stop context-switching to manage your distributed SQL infra

I remember the old days of manual scaling. You'd jump into a CLI, check your metrics, realize you needed another node or a capacity adjustment, log into a web console, navigate three layers deep into ...