Amazing, exactly what Workers lacked, I was quite annoyed that a worker would spin up when 99% of my requests then return a cache. Waste of 3ms, times millions.