Software doesn’t scale through architecture and automation alone. New, more difficult problems appear that didn’t exist before, causing new downstream consequences. Something interesting happens when you run more than 1,000 servers, as we do at WP Engine1, powering hundreds of thousands of websites. 1 Editor’s Note: As of 2024, this article is ten years old; we now run twenty times as many servers, and the lesson of this article continues to be accurate. Suppose that on average a server experien...