Every DNS failure has a leg where it broke. These are live zones on a real grid, each misbehaving in exactly one documented way, so you can dig at them and see the difference between "the forwarder is down" and "the delegation is lame" — the two that look identical from the client seat.
REFUSED, SERVFAIL, NODATA, black-hole, truncation, slow. One zone per fault. Dig each and watch NIOS reshape the answer the client actually sees.
content nextA healthy delegation, a lame one into TEST-NET, and the query-type split that tells them apart in dnstap. The distinction the whole thesis rests on.
content nextA zone that answers and holds nothing anyone has queried since 2019 — config that looks like a path and resolves nothing. The real-world trap.
content nextNIOS changes defaults between releases and the change notes are scattered. GoldenGrid runs all five and diffs them, so "when did this default change, and to what" is a table you can read.
| Version | Template | Defaults captured |
|---|---|---|
| 8.5.2 | nios-852-tmpl | — pending capture |
| 9.0.3 | nios-903-tmpl | — pending capture |
| 9.0.7 | nios-907-tmpl | — pending capture |
| 9.0.8 | nios-908-tmpl | — pending capture |
| 9.1.0 | nios-910-tmpl | — pending capture |
This grid manages itself the way a real one should. The admin password is rotated on a schedule by WAPI — not a static credential pasted in a doc. The rotation is the exhibit: grid hygiene, demonstrated, not described.