node-craft-ops
/node-craft-ops
Porządkuje operacje backendu Node.js i TypeScript: wdrożenia, sondy zdrowia, graceful shutdown, kopie zapasowe PostgreSQL i CI/CD.
- Autor
- hasaneks ↗
- Repozytorium
- node-craft ↗
- Licencja
- MIT
- Ostatnia zmiana repozytorium
- 04.11.2026
Ocena i bezpieczeństwo
- Pliki w paczce
- 6
- Skrypty
- 0
- Komentarze HTML
- 0
- Adresy URL w SKILL.md
- 0
Brak sygnałów. Repozytorium zmienione 04.11.2026, bez odwołań do wycofanych modeli i API.
MIT, otwarta: można używać i zmieniać, z zachowaniem informacji o autorze.
Co robi i kiedy
Porządkuje operacje backendu Node.js i TypeScript: wdrożenia, sondy zdrowia, graceful shutdown, kopie zapasowe PostgreSQL i CI/CD.
Oryginalny opis (EN)
Operations rules for a Node.js/TypeScript backend: deploying on any platform (PaaS, containers or VMs: build, start, a single release step for migrations, port and host binding, private networking, health checks, instances, draining and graceful shutdown, versioned service config), liveness and readiness endpoints, CI/CD quality gates from immutable install to staging smoke test, migration release and rollback strategy, PostgreSQL backup, restore tests, RPO and RTO, and runbooks for alerts, failed jobs, webhook replay and incidents. Use when setting up or changing deployment, CI, health checks
Jak działa
- Rules that apply everywhere
- Release checklist
Podgląd SKILL.md
--- name: node-craft-ops description: "Operations rules for a Node.js/TypeScript backend: deploying on any platform (PaaS, containers or VMs: build, start, a single release step for migrations, port and host binding, private networking, health checks, instances, draining and graceful shutdown, versioned service config), liveness and readiness endpoints, CI/CD quality gates from immutable install to staging smoke test, migration release and rollback strategy, PostgreSQL backup, restore tests, RPO and RTO, and runbooks for alerts, failed jobs, webhook replay and incidents. Use when setting up or changing deployment, CI, health checks, shutdown, backups or runbooks, or preparing a release. Never starts paid resources or a production release on its own." --- # Ship and run the backend Follow `node-craft` for the working rules. Writing a deploy configuration is not a request to deploy: creating paid resources, changing production settings and releasing to production need explicit approval each time. | Area | Read | |---|---| | Choosing and recording the platform, build, start, release step, networking, instances, storage, config | `references/deployment.md` | | `/health/live`, `/health/ready`, signals, draining, crash handling | `references/health-shutdown.md` | | CI gates, release flow, migration and rollback order | `references/ci-gates.md` | | Backups, restore tests, RPO/RTO, data recovery | `references/backup-restore.md` | | Alerts, failed jobs, webhook replay, incident steps | `references/runbooks.md` | ## Rules that apply everywhere - **Same versions everywhere:** Node and Yarn from project metadata in local, CI and deploy builds. - **Build builds; it does not migrate.** Migrations run once per release in a single release step, with a lock; never at every replica start; never inside the image build. - **A failed migration stops the release.** No traffic goes to a version whose migration failed. - **Stateless processes.** Sessions, locks, rate limits and schedules that must be shared live in PostgreSQL or Redis, not in process memory. User files live in object storage. - **Health checks tell the truth.** Never make a health endpoint return success to get past a deploy. - **Shutdown is bounded.** Stop accepting work, finish in-flight work within the platform's draining time, close connections, exit. - **Recovery is tested.** A backup that was never restored is a hope. Record restore tests with dates and durations. - **No platform is assumed.** Use the project's platform, or record the choice in an ADR; map its pieces (release step, health checks, grace period, networking) in `STACK.md` (`references/deployment.md`). Platform features, limits, prices and config formats change; check the current docs before relying on them. ## Release checklist 1. CI green on the commit (`references/ci-gates.md`). 2. Migration reviewed for lock time and old-code compatibility; backup recent; data-loss risk approved if any. 3. Staging deploy: release-step migration succeeds, health checks pass, smoke tests pass, sandbox integrations checked. 4. Production release with approval; watch errors, latency, queue and outbox lag for the agreed period. 5. Rollback plan known: application rollback, and which applies to the database (old code compatible, forward fix, or restore). When a release or migration fails: stop, report the failure, its impact and the state it left openly, then choose forward fix or restore (`references/backup-restore.md`). Never hide it by faking health, skipping a gate or rerunning until it passes.
Pliki w paczce (6)
- SKILL.md 3 KB
- references/backup-restore.md 2 KB
- references/ci-gates.md 3 KB
- references/deployment.md 6 KB
- references/health-shutdown.md 3 KB
- references/runbooks.md 2 KB
Wersje
W bazie nie ma innych skilli o nazwie node-craft-ops.
Migawka bazy z 13.11.2026 · wersja paczki 79975a7901f5 (skrót SHA-256 całej paczki)
Pobrania w ostatnich 30 dniach: 0 · ranking
Zgłoś problem z tym skillem: szkodliwa treść, błędny opis albo prośba autora o usunięcie.
Podobne skille
- 100health-checks
Opisuje wzorce kontroli kondycji i łagodnego zamykania: sondy liveness i readiness, kontrole zależności, sondy Kubernetes, Spring Actuator i zamykanie bez przestojów. - 100backup-drill
Sprawdza, czy kopie zapasowe faktycznie się odtwarzają: przywraca dane na cel testowy, mierzy RPO i RTO i pisze runbook. - 100data-integrity
Chroni integralność i odzyskiwalność danych: RPO, RTO, kopie zapasowe a migawki, testy przywracania i zgodność odczytu z zapisem. - 100db-backup
Konfiguruje automatyczne kopie zapasowe, odtwarzanie i procedury awaryjne dla PostgreSQL, MySQL, MongoDB i Redis: harmonogramy, odtwarzanie do punktu w czasie, replikacja między regionami i runbooki. - 90resiliency-recovery
Buduje lub audytuje dostępność i odtwarzalność systemu rządowego w Singapurze: wielostrefowość, kopie zapasowe, testy odtwarzania i RTO/RPO. - 100disaster-recovery
Ustala standardy kopii zapasowych i odtwarzania: RTO, RPO, zasadę 3-2-1, niezmienne kopie, przywracanie testowe i monitoring kopii.