Because the absence of failures so far is a statement about age, not about reliability. Hardware failure rates are low in the first few years and rise afterwards, so a young estate has not demonstrated anything yet — it has simply not reached the part of its life where components fail.

That the first one is handled by people who have never handled one. Nobody has opened the chassis, nobody knows which spare fits, nobody has tested whether the backup actually restores, and the array rebuild that has never been observed is happening for the first time under pressure. The hardware event is usually survivable; the inexperience is what extends it.

Rehearse the parts of it that are safe to rehearse. Pull a redundant power supply on a machine that has two and watch what the system does. Fail a mirrored drive deliberately and let it rebuild. Restore a backup to somewhere harmless and confirm it works. All three can be done without risk and all three answer questions you would otherwise answer during an incident.

No, though that is where our depth is. For a young estate the more useful thing we provide is the matched spare for the platform you have, identified from its serial so it is certain to fit.