Structured cabling + cable lifecycle
A cable just died. Not pulled, not unplugged -- it gave up. Cables age, get stepped on, and fail. Anything downstream of that cable is offline until you yank it and lay a fresh one. The fix is cheap; the outage isn't.
A tidy rack -- permanent runs on trays, churn on short patch-panel jumpers -- is a rack where the next cable failure is one obvious replacement, not an archaeology dig. Loose floor snakes get trodden on and bent past their limit; supported, labelled runs last their rated life. Good cabling discipline is the difference between a five-minute swap and a 3am scavenger hunt.
Detailed explanation
Cables carry a wear curve driven by age + power-on hours + bend-radius violations. CableSevered fires when the wear roll hits -- affected ports cycle to Down, reachability sweeps re-evaluate, and any host that drops off the fabric strands its VMs. Recovery is a RemoveCable + fresh LayCable; bonded LAG members degrade rather than fail outright, which is half the reason to bond in the first place.
Bend-radius violations are a direct wear input, so a route that snakes around a tight corner ages faster -- which is exactly why the planner enforces radius on both in-rack strip corners and tray / J-hook chains. Structured cabling (panels + trays) keeps runs inside their rated radius and off the floor, flattening the wear curve. Replace anything past its rated service life on a maintenance window, not after it broke. The engine's CablingDiscipline metric tracks how much of your plant is structured versus ad-hoc.