**Parent Topic**: [[Software/README]] ## Two Views Debois's opening lesson from running a single web server: there are always two views — the **technician view** and the **end-user/business view**. "Monitoring isn't about setting up a system. It's there to support the business, to make sure they can do their work." For a website, that ability to work is rephrased as keeping the site available. ## The User Always Beats Your Monitoring Every escalation in his stories was a problem an end user detected before the monitoring did — because "the user will always be running the business and transactions all the time," exercising input combinations no test covers. If you have many users, turn this to your advantage: usage statistics and help-desk call volume become **indirect failure indicators**, catching problems your technical checks miss. ## Monitoring Is a Journey The second lesson: monitoring starts simple and improves along the way — it took Debois years to grasp its full potential. Don't expect a finished monitoring system; expect to iterate. This framing underlies all three steps, beginning with [[Map Dependencies Before Monitoring]]. --- *Source: [[Web Operations]] (Allspaw & Robbins, O'Reilly 2010) — Ch 6 — Monitoring* ## Related - [[Failure Modes Mirror Above and Below the Line]] — Systems thinking parallel: failure patterns in technical systems mirror those in human coordination layers - [[Jevons Paradox as AI Skill Leverage Crescendo]] — Economic paradox: efficiency gains increase consumption, not reduce it — better monitoring tools should enable more monitoring activity - [[Risk Aversion as the Hidden Cost of Success — How Reliability Requirements Lock In Expensive Solutions]] — Monitoring systems risk becoming overly conservative and expensive as reliability culture locks in - [[Quality Degradation as Hidden Hyperinflation Signal]] — Quality-value tension: monitoring can become more complex while delivering less business value