Up isn't the same as working.

The report says 99.95%. The branch manager says the tills kept freezing all afternoon. Both are telling the truth. Availability measures one thing: whether there is a connection. It does not measure whether the connection carries the work. A link can sit inside its SLA all month and still drop packets, spike latency and stall the exact applications a site runs on. None of that shows up as downtime. The circuit was never off. It just wasn’t working. That gap, between available and working, is where the cost hides. The payment that failed. The video call that broke. The shift that lost an hour. No outage was ever logged, because by the only number anyone measured, nothing went wrong. This is why “up” is a comfortable number and the wrong one to trust. It is easy to hit and easy to report. It also tells you almost nothing about the experience of the people relying on the network. The number that matters is different. Not “is the circuit up?” but “is it carrying what the business needs, right now, measured from where the work happens?” That means watching the things applications actually feel, latency, loss, jitter, throughput against the SLA, continuously, and reporting them plainly. Up isn’t the same as working. Measure the second one.