← ClaudeAtlas

ingress-migration-advisorlisted

Read-only planner that inventories every Ingress in a Kustomize repo, scores each service on five migration-readiness dimensions, and recommends one of four paths per service (direct-gateway, two-step, swap-only, defer). Output is a Mermaid Gantt plan plus ready-to-paste Zeus commands. Critical traffic-tier services are vetoed to defer. Services already on Traefik Ingress (sourceClass=traefik) auto-route to direct-gateway. Never mutates the repo; produces docs/reports/ ingress-migration-advisor/<slug>/plan.md and state.yaml. Use for end-of-life planning (ingress-nginx EOL 2025), migration sequencing, or per-service path advisory. Requires docs/ingress-tier-map.yaml in the consumer repo.
qwedsazxc78/devops-ai-skill · ★ 2 · DevOps & Infrastructure · score 68
Install: claude install-skill qwedsazxc78/devops-ai-skill
# ingress-migration-advisor Skill Invoked by Zeus pipeline `*ingress-migration-advisor`. This skill sits **upstream** of the three migration skills (`nginx-to-traefik`, `gateway-api-migration`, `nginx-to-gateway`). It does not invoke them — it produces a recommendation plan, and the operator runs each batch through Zeus manually using the suggested commands. Read this skill in full before answering questions about "which path should I migrate service X on" or "what is our EOL migration plan". ## Canonical references | File | When to read | |---|---| | `references/scoring-model.md` | Step 2 — the five-dimension rubric + critical-tier veto rule | | `references/decision-matrix.md` | Step 3 — score-band → path mapping + sourceClass shortcut | | `references/plan-template.md` | Step 5 — Mermaid Gantt + per-service table + per-batch command block | ## Bundled scripts | Script | Used by | Purpose | |---|---|---| | `scripts/inventory_all_ingresses.py` | Step 1 | Build every overlay, classify rendered Ingresses (reuses classify_ingress.py from gateway-api-migration) | | `scripts/score_services.py` | Step 2, 3 | Apply veto + rubric, decide path per service, emit decisions | | `scripts/render_plan.py` | Step 5 | Substitute plan-template.md with state.yaml data | The scripts produce structured JSON / YAML. The model's job is to invoke them and weave their output into the report — not to re-derive scoring or decision logic in natural language. ## Activation Triggered explicitly b