cloud-and-devops-architect
Architect-level cloud direction — choosing provider services from operational reality, environment and account topology, infrastructure-as-code ownership, network and identity boundaries, reliability targets, and the cost model a system is designed against. Use before any provider, orchestration or IaC skill.
- License 许可
- MIT
- 最近更新
- 2026-11-13
- Stars
- ★ 0
- Forks
- 0
01 前言 Front Matter
- name
- cloud-and-devops-architect
- license
- MIT
- metadata
- { "crusor-path": "cloud-and-devops-architect", "crusor-domain": "cloud-and-devops-architect", "crusor-origin": "original", "crusor-sources": "12factor.net; aws.amazon.com/architecture/well-architected; kubernetes.io/docs; developer.hashicorp.com/terraform", "crusor-triggers": "cloud architecture, infrastructure, devops, iac, kubernetes, terraform, environments, networking, cloud cost", "crusor-verified-on": "2026-10-30" }
- description
- Architect-level cloud direction — choosing provider services from operational reality, environment and account topology, infrastructure-as-code ownership, network and identity boundaries, reliability targets, and the cost model a system is designed against. Use before any provider, orchestration or IaC skill.
- compatibility
- AWS/GCP/Azure; Kubernetes; Terraform or Pulumi
03 SKILL.md 正文
Cloud & DevOps Architect
Cloud architecture is renting failure modes. Every managed service you adopt buys operational simplicity and sells you a new dependency, a new limit, and a new bill.
When to use
- Standing up or reviewing a cloud estate, or choosing between providers.
- A cost, reliability or identity problem that has reached architecture level.
- Any plan that touches networking, IaC, or managed-service selection.
When NOT to use
- Application-level design questions — that is the website, mobile or data domain.
- A single-service tweak with no cross-cutting consequence.
Leaf map
Seven shelves, 35 books.
| Shelf | Books |
|---|---|
providers-architect/ |
aws · gcp · azure · cloudflare |
identity-architect/ |
iam · multi-account · secrets-and-keys |
networking-architect/ |
vpc-networking · dns-and-edge · network-security |
compute-architect/ |
kubernetes · serverless · ec2 · ecs · lambda · autoscaling |
data-services-architect/ |
object-storage · managed-databases · messaging · caching |
iac-architect/ |
terraform · pulumi · cloudformation · ansible |
platform-architect/ |
helm · gitops · service-mesh · kubernetes-security · platform-engineering |
operations-architect/ |
reliability · cost-optimization · sre-observability · chaos-engineering · backup-and-dr · cloud-incident-response |
Read this crown, then take the one or two books the decision needs. The shelf tells you the axis; the book tells you the trade-off.
Decision 1 — Topology before services
accounts/subscriptions per environment (dev, staging, prod, shared)
network per environment (VPC/VNet), private by default
workload per bounded context
data per owner — never one shared database for everything
- Environments are isolated by account, not by naming convention.
- Blast radius is a design input: what fails together should be grouped together, and production must not share a control plane with staging.
- One region until a documented requirement forces multi-region; multi-region multiplies every consistency decision.
Decision 2 — Managed vs self-run
| Situation | Choose |
|---|---|
| commodity (queues, object storage, DNS) | managed |
| undifferentiated but operationally risky (Kafka, Postgres at scale) | managed first |
| core differentiator, hard constraints | self-run, with an owner |
- Every self-run component needs an owner, a runbook, and an upgrade path.
- Managed services still need limits knowledge: quotas, throttles, regional availability, and the pricing cliff.
Decision 3 — Infrastructure as code, always
- No console changes in production; drift is detected and reverted.
- State is remote, locked, and per environment; state contains secrets, so it is access-controlled.
- Modules are versioned and consumed by reference, not copy-paste.
- Plan output is reviewed like code; apply happens in the pipeline.
Decision 4 — Identity and network boundaries
- Workloads authenticate with roles/workload identity, never static keys.
- Least privilege per service: each has its own identity and its own policy.
- Private subnets for data and internal services; public only where required (load balancer, CDN edge).
- Egress is controlled: an allow-list beats an open NAT for both security and cost.
- Secrets in a managed secret store with rotation (see
secrets-management-architect).
Decision 5 — Reliability targets
- Define SLIs/SLOs per user journey before choosing redundancy.
- Single-AZ by default, multi-AZ for anything with an SLO above 99.9%.
- Backups: automated, encrypted, restore-tested; RPO/RTO written down.
- Chaos is scheduled: kill an instance, drain a zone, expire a credential.
Decision 6 — Cost is architecture
- Tag everything (service, environment, owner) from day one; untagged spend cannot be attributed.
- Budgets and alerts per environment; anomaly detection on.
- The expensive patterns are predictable: idle clusters, cross-AZ chatter, unbounded logs, over-provisioned instances, storage never deleted.
Failure modes at scale
- Console changes that IaC later reverts or duplicates.
- One account for everything; a staging mistake reaches production credentials.
- Static access keys in CI.
- Multi-region "for safety" with no failover test.
- Costs discovered at the end of the quarter, untagged and unattributable.
Verification receipts
- Account/environment diagram with network and identity boundaries.
- IaC plan/drift report clean; no manual production changes in the last period.
- IAM review: per-workload roles, no long-lived keys.
- Restore drill with measured RTO; failover drill if multi-AZ/region claimed.
- Cost report by tag with budget alerts active.
04 同仓库更多 Skills
interaction-patterns-architect
Architect-level interaction design in code — choosing established patterns over invented ones, feedback and latency behaviour, destructive-action safeguards, navigation and overlay conventions, and the state model behind every interactive surface. Use when a screen has to behave, not just look.
design-system-ops-architect
Architect-level operation of a design system in code — token pipelines and versioning, component API and variant policy, the catalogue as the contract, adoption metrics, and the governance that stops a system from forking into three. Use when a component library must survive several teams.
design-handoff-architect
Architect-level design handoff — what must be specified before implementation starts, tokens over raw values, responsive and content-limit contracts, state matrices, and the review loop that catches drift between design and code. Use when a design is about to become code.
content-microcopy-architect
Architect-level UI content — where strings live, voice and terminology consistency, error and empty-state copy as part of the interface, translation-ready authoring, and the content model a product needs before it multiplies screens. Use when the words in the interface matter as much as the pixels.
ui-ux-implementation-architect
Architect-level direction for turning a design into working UI — the handoff contract, component ownership versus screen-level layout, interaction and state conventions, design QA gates, and the review habits that keep an implementation faithful without frozen pixels. Use before any specific UI implementation skill.
test-strategy-architect
Architect-level test strategy — allocating assertions across unit, integration, contract and end-to-end layers, designing for testability through dependency boundaries, fixture and data strategy, and the flake and coverage policies that keep a suite trustworthy. Use when defining how a codebase is tested.