docs(failure-domains): name domains with labels instead of integer ids

[schmidt-scaled]

Documents the failure-domain / physical labels feature built on simplyblock/sbcli branch feature/fd-labels (commit d9732101).

What changes for the operator

sn add-node --failure-domain takes a label — RACK1, AZ2, HOST1 — instead of a non-negative integer. The integer is still the internal identity (placement, the distrib cluster map and the expansion planner all key off it, and the v2 API keeps its integer failure_domain field, now alongside failure_domain_label), but operators no longer invent or track it.

Pages touched

• architecture/concepts/failure-domains.md — domains are named rather than numbered; the label→id mapping is described as an internal detail.
• non-kubernetes/operations/failure-domains.md — examples switched to RACK1/RACK2; two new sections:
  • Label Syntax — a letter followed by up to 31 of [A-Z0-9_-], case-insensitive, stored upper-cased.
  • Labels on Existing Clusters — initialization through cluster update, which names each id in service FD<id> / HOST<id>. Idempotent, so a later rename survives; refuses to take a derived name already owned by a different id.

Compatibility, documented explicitly

An all-digits --failure-domain value is still read as the internal id, so existing scripts, CI bootstraps and the Kubernetes operator keep working unchanged. The node list's Failure Domain column shows the label and falls back to the id for a cluster that has not been initialized.

Notes for review

• Base is main: this repo has no dev branch (default is main, with releases cut to release/* / R25.*). Happy to retarget.
• reference/operator/reference.md is generated CRD documentation and is untouched — the Kubernetes failureDomain field lives in the operator repo and would need its own change to accept labels.
• Ships with the sbcli branch; merge order should follow whenever that lands.

🤖 Generated with Claude Code

[github-actions]

🤖: Deployment available as https://docstest.simplyblock.io/fd-labels