[Epic]: Platform reliability debt — clear the P1 defects found in the 2026-08-13 cluster audit

## Summary
A read-only audit of the Hetzner `home-remote` cluster on 2026-08-13 found no P0 issues, but four P1 degradations and two hygiene items — four of which had no tracking issue at all. This epic groups the six follow-up Stories needed to clear that debt.

## Intent / Source of truth
2026-08-13 cross-repo backlog consolidation + live cluster audit (this epic's Stories cite the specific live evidence: `kubectl get certificate`, ArgoCD sync status, Job/Pod state, CronJob history, and `charts/mcp` config).

## Scope
- [ ] Six child Stories (linked as sub-issues) each independently verifiable and closeable
- [ ] No epic-level deliverable beyond tracking; each Story carries its own verification

## Out of scope
- Any P2/P3 items surfaced by the same audit that are not P1 — those are tracked separately or accepted as-is
- New feature work unrelated to the audit findings

## Verification
All six linked Stories closed with their individual verification steps passing on the live cluster.

## Risk assessment
Low — this is a tracking epic. Risk lives in the individual Stories (see each).

## AI Usage Declaration
Drafted with AI assistance during the 2026-08-13 cross-repo backlog consolidation. A human owns intent, verification and consequences; acceptance criteria must be verified against the live system before this is closed.

This is a tracking epic rather than an implementation task. Begin with the six linked child Stories and their cited evidence—kubectl get certificate, ArgoCD sync status, Job/Pod state, CronJob history, and charts/mcp config—then follow each Story's verification steps. Done means all six Stories are closed and their live-cluster checks pass.

索引モデルが issue の本文から書いたものです。