[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7AM8Z43WADN4SHB0V0ZM7B3":3,"skill-same:01M7AM8Z43WADN4SHB0V0ZM7B3":48},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":29,"ownerHtmlUrl":30,"ownerBlogUrl":31,"ownerDescription":32,"updatedAt":33,"install":34,"resources":47},"01M7AM8Z43WADN4SHB0V0ZM7B3","cloud-reference",null,"Use when checking a cloud anti-pattern, weighing a contested question, confirming whether a pricing, regulatory or provider claim is still current (snapshot verified August 2026), finding the primary sources worth reading instead of blogs, or needing the design checklist, the numbers worth remembering, and a triage list. Companion to the other cloud-computing skills.","the-vibey-project\u002Fvibey","the-vibey-project","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F329298758?v=4","01M7AJHGZ9722CTJ4JNQZVRH0Q","the vibey project","01M7AJHGZVMG2K7TGXQ6XFHVG3",5,"MIT","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fcloud-computing\u002Fskills\u002Fcloud-reference\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: cloud-reference\ndescription: \"Use when checking a cloud anti-pattern, weighing a contested question, confirming whether a pricing, regulatory or provider claim is still current (snapshot verified August 2026), finding the primary sources worth reading instead of blogs, or needing the design checklist, the numbers worth remembering, and a triage list. Companion to the other cloud-computing skills.\"\n---\n\n# Cloud Computing: Anti-Patterns, Contested Questions, Currency, and Canon\n\n> **Part 5 of 5** of the *Cloud Computing* reference (plugin `cloud-computing`), covering §15–§20. Sibling skills: `cloud-models-providers-and-primitives` (§0–§3), `cloud-architecture-and-resilience` (§4–§6), `cloud-cost-security-and-operations` (§7–§9), `cloud-migration-sovereignty-and-ai-workloads` (§10–§14). Section numbers are shared across the set; a reference written as §N → `skill` points into that sibling skill.\n>\n> **Currency:** Verified August 2026. See §17 below for the currency snapshot and what goes stale first.\n\n> **How to read this.** Reference, not a certification guide. Three markers:\n> - **[DURABLE]** — architectural principles and trade-offs that survive provider churn.\n> - **[VERSIONED]** — market data, pricing, service names, regulation. **Verify before\n>   quoting.**\n> - **[CONTESTED]** — genuine disagreement.\n>\n> **⚠️ GOTCHA** boxes mark the mistakes that produce outages or surprise invoices.\n>\n> **Provider-neutral by default**, with AWS\u002FAzure\u002FGCP names where the concept needs an\n> anchor. Service names change; the primitives don't.\n>\n> **The three framings that organize everything below:**\n> 1. **The cloud is someone else's computer, rented, with an API and a metered bill.**\n>    Every genuine advantage (elasticity, global reach, managed services, capex→opex) and\n>    every genuine failure mode (cost surprise, lock-in, shared fate, opaque outages)\n>    follows from that one sentence.\n> 2. **⚠️ Your failure domain is always larger than your architecture diagram says.**\n>    The 2025 outages taught this expensively: multi-AZ deployments failed because of\n>    control-plane dependencies nobody had drawn (§6 → `cloud-architecture-and-resilience`). **Map what you actually depend on,\n>    not what you think you depend on.**\n> 3. **Cost is an architectural property, not an operational afterthought.** It is\n>    determined by design decisions — data placement, service selection, egress paths —\n>    and **discovered on an invoice 30 days later** (§7 → `cloud-cost-security-and-operations`). Design for it up front or pay\n>    for it forever.\n\n---\n\n## §15. Anti-Patterns\n\n| Anti-pattern | Why |\n|---|---|\n| Lift-and-shift everything, then be surprised by the bill | Fixed-capacity design on per-hour billing (§10 → `cloud-migration-sovereignty-and-ai-workloads`) |\n| Treating cost as an ops problem | ⚠️ **It's set at design time** (§7 → `cloud-cost-security-and-operations`) |\n| No tagging strategy | **You can't allocate what you can't attribute** (§7 → `cloud-cost-security-and-operations`) |\n| Ignoring commitment discounts | **Fewer than half of orgs use any** — free money (§7.1 → `cloud-cost-security-and-operations`) |\n| Never using spot for interruptible work | 60–90% cheaper (§3 → `cloud-models-providers-and-primitives`, §13 → `cloud-migration-sovereignty-and-ai-workloads`) |\n| Provisioned block storage sized for peak, forever | You pay for the ask, not the use (§3 → `cloud-models-providers-and-primitives`) |\n| No storage lifecycle policy | Minutes to set, runs forever (§3 → `cloud-models-providers-and-primitives`) |\n| **Assuming multi-AZ protects against regional failure** | ⚠️ **The single most expensive lesson of 2025** (§6 → `cloud-architecture-and-resilience`) |\n| Multi-region compute with single-region data | The dependency that actually took people down (§6.2 → `cloud-architecture-and-resilience`) |\n| Not auditing whether failover routes through us-east-1 | Global services anchor there (§6.2 → `cloud-architecture-and-resilience`) |\n| Monitoring and alerting behind the same provider you're monitoring | ⚠️ **Failed exactly this way in 2025** (§6.2 → `cloud-architecture-and-resilience`) |\n| CI\u002FCD and IaC state in the region you'd fail over *from* | Can't deploy the fix (§6.3 → `cloud-architecture-and-resilience`) |\n| Untested failover runbook | It's a hypothesis, not a plan (§6.3 → `cloud-architecture-and-resilience`) |\n| Binary up\u002Fdown with no degraded mode | Read-only beats down, and costs far less (§6.3 → `cloud-architecture-and-resilience`) |\n| Long-lived access keys | ⚠️ **The most common breach vector.** Use workload identity (§8 → `cloud-cost-security-and-operations`) |\n| Wildcard IAM policies that were \"temporary\" | They never are (§8 → `cloud-cost-security-and-operations`) |\n| Treating a VPC as a trust boundary | Network location ≠ authorization (§8 → `cloud-cost-security-and-operations`) |\n| Assuming provider compliance certification = your compliance | It enables it; config is yours (§8 → `cloud-cost-security-and-operations`) |\n| Alerting on CPU instead of user-visible symptoms | Alert fatigue is a reliability problem (§9 → `cloud-cost-security-and-operations`) |\n| **Assuming data residency equals sovereignty** | ⚠️ **The \"residency illusion\"** — CLOUD Act reach (§11.1 → `cloud-migration-sovereignty-and-ai-workloads`) |\n| Signing or auto-renewing EU cloud contracts without the 2027 ban in view | ⚠️ **Locks in an outdated fee structure** (§11.2 → `cloud-migration-sovereignty-and-ai-workloads`) |\n| Thinking egress fees were the whole of lock-in | APIs, formats and IAM outlast the invoice line (§11.2 → `cloud-migration-sovereignty-and-ai-workloads`, §12 → `cloud-migration-sovereignty-and-ai-workloads`) |\n| Chatty service mesh spanning AZs | Cross-AZ transfer is a real, invisible cost (§11.3 → `cloud-migration-sovereignty-and-ai-workloads`) |\n| Kubernetes for three services and five engineers | ⚠️ **Adopting a platform team's problem without the platform team** (§5 → `cloud-architecture-and-resilience`) |\n| Serverless for sustained high-volume load | The cost curve inverts (§5 → `cloud-architecture-and-resilience`) |\n| Multi-cloud portability as a default goal | Lowest common denominator, multiplied ops (§12 → `cloud-migration-sovereignty-and-ai-workloads`) |\n| Accidental lock-in | ⚠️ **You paid the abstraction cost and got locked in anyway** (§12 → `cloud-migration-sovereignty-and-ai-workloads`) |\n| Statically provisioned GPU fleets | **30–40% utilization; fastest-growing waste** (§13 → `cloud-migration-sovereignty-and-ai-workloads`) |\n| Training on on-demand instead of spot with checkpointing | Bursty and interruption-tolerant by nature (§13 → `cloud-migration-sovereignty-and-ai-workloads`) |\n| Retrofitting AI cost attribution later | **What the 2026 waste reversal is measuring** (§7.1 → `cloud-cost-security-and-operations`, §13 → `cloud-migration-sovereignty-and-ai-workloads`) |\n| Quoting a cloud market share figure without its source | ⚠️ **They differ by 4+ points on methodology** (§2 → `cloud-models-providers-and-primitives`) |\n\n---\n\n## §16. Contested Questions\n\n**16.1 Is cloud cheaper than on-prem?** *For*: no capex, elastic, no datacenter staff,\nyou pay for what you use. *Against*: at **steady, predictable, high utilization**, owned\nhardware is frequently cheaper — and that's exactly the profile that gets repatriated.\n**[CONTESTED, and the honest answer is workload-dependent.]** The 2026 data supports a\nmiddle reading: **~21% of workloads repatriated, but new cloud workloads outstrip the\nexits.** ⚠️ **From 2027 the EU removes exit-cost distortion from this comparison, which\nmeans the TCO math should genuinely be re-run** rather than assumed.\n\n**16.2 Is repatriation a real trend or vendor noise?** Real but bounded. **Selective\nrepatriation of steady-state, high-utilization workloads is happening**, driven partly by\nFinOps tooling finally making the comparison visible. **Wholesale cloud exit is not.**\n\n**16.3 Multi-cloud: insurance or expensive complexity?** §12 → `cloud-migration-sovereignty-and-ai-workloads`. The 2025 outages strengthened\nthe resilience argument; the operational reality strengthened the complexity argument.\n**Both got more true at once.**\n\n**16.4 Is Kubernetes over-adopted?** **[CONTESTED but leaning yes.]** Excellent at scale\nwith a platform team, poor fit below that, and adopted for reasons that are often about\nrésumés and perceived sophistication rather than requirements.\n\n**16.5 Does serverless deliver?** *For*: genuinely excellent for event-driven, spiky, and\nlow-volume workloads; operational burden near zero. *Against*: cost inversion at sustained\nvolume, cold starts, testing friction, vendor coupling. **The pendulum settled around\n\"right tool for specific shapes\" rather than a default.**\n\n**16.6 Is sovereign cloud meaningful or theatre?** *For*: the CLOUD Act tension is real\nand residency genuinely doesn't solve it. *Against*: much of what's marketed as sovereign\nis residency with extra branding. **⚠️ The EU's proposed Cloud and AI Development Act\n(June 2026) is an attempt to settle this by defining graded sovereignty levels** — which\nis itself an admission that the term is currently unregulated marketing.\n\n**16.7 Are the hyperscalers over-building for AI?** ⚠️ **Genuinely unknown, and the stakes\nare large.** ~$570B+ of combined 2026 capex guidance is a bet that AI demand\nmaterializes. Revenue growth currently supports it. **If it disappoints, the write-downs\nwould be historically large** — and buyers should factor that uncertainty into\nlong-horizon commitments.\n\n---\n\n## §17. Currency Snapshot — verified August 2026\n\n| Thing | Status as of Aug 2026 | Decay risk |\n|---|---|---|\n| **Market structure** | Q1 2026 cloud infrastructure spend **~$129B, +35% YoY**, ninth consecutive quarter of accelerating YoY growth. Big Three **~60–68%** depending on methodology. ⚠️ **Synergy: AWS 28% \u002F Azure 21% \u002F GCP 14%. Other sources: 30\u002F24–25\u002F13 or 32\u002F23\u002F11** — **cite source and scope** (§2 → `cloud-models-providers-and-primitives`). Azure and GCP growing faster off smaller bases; **Microsoft doesn't break out Azure revenue, only growth rate** | **High** |\n| **Capex** | 2026 guidance: **Amazon ~$200B, Microsoft ~$190B, Alphabet ~$180–190B.** Guidance, not actuals | **High** |\n| **AI share of cloud spend** | **~19% in 2026, up from ~8% in 2023** | **High** |\n| **⚠️ Cloud waste** | **Flexera 2026: 29% of IaaS\u002FPaaS wasted, up from 27% — reversing a five-year decline**, attributed to **AI adoption outpacing tagging\u002Fattribution\u002Fgovernance**. n=753. **Managing cloud spend now the #1 priority (84%), above security** | Medium |\n| **FinOps adoption** | **FinOps teams ~63%, CCoE ~71%.** ⚠️ **Fewer than half use any one commitment discount per provider.** **AI cost management the top future priority (98%)** and most-wanted skillset. **FOCUS** standardizing cross-provider cost data. Practitioners report **the easy waste is gone** — remaining savings are smaller and harder | Medium |\n| **Repatriation** | **~21% of workloads repatriated**, but **new cloud workloads outstrip exits — cloud continues to grow** | Medium |\n| **GPU economics** | **GPUs ~18% of spend at AI-forward orgs (up from ~4% in 2023); statically provisioned fleets at 30–40% utilization.** Idle accelerators = fastest-growing waste category | **High** |\n| **⚠️ Oct–Nov 2025 outages** | **AWS us-east-1, 20 Oct, ~14–15h** — latent **race condition in DynamoDB's DNS management** wiped DNS records, cascading across dozens of services. **Azure, 29 Oct** — **Azure Front Door** \u002F global identity. **Cloudflare, 18 Nov** — Bot Management config **doubled a feature file past a hard-coded limit**, crashing proxies globally; **engineers spent ~2h believing it was a DDoS**. Common pattern: **subtle single-subsystem defect → global cascade**; diagnosis is **immature blast-radius modelling**, not human error | Low (historical) |\n| **us-east-1 concentration** | Reported at **~69% of AWS usage** by one analysis and **44%+ of AWS requests** by another. **Global operations (parts of IAM, Route 53) anchor there** | Medium |\n| **⚠️ EU Data Act** | Reg. (EU) 2023\u002F2854. In force 11 Jan 2024. **Switching rights enforceable 12 Sept 2025.** **12 Sept 2026: products released after must have user data accessible by default.** **⚠️ 12 January 2027: all switching charges including egress fees banned outright** — IaaS, PaaS and SaaS, **for every provider serving EU customers**. Transition allows only direct, transparent, pre-agreed cost. Enforcement by **national authorities**, penalties set in **local legislation**. AWS\u002FAzure\u002FGoogle waived full-exit egress in 2024 (narrower than 2027 requires) | **High** |\n| **DMA \u002F sovereignty** | **DMA investigating AWS and Azure as potential gatekeepers**; ⚠️ **analysts flag genuine friction with the Data Act** — conflicting timelines, split enforcement (member states vs. Commission). **3 June 2026: Commission proposed the Cloud and AI Development Act** (Tech Sovereignty Package) to define **graded sovereignty levels** — ⚠️ **not yet passed; details will change** | **High** |\n| **Sustainability** | Well-Architected pillar; **>half of orgs have or plan a sustainability initiative including cloud carbon tracking** | Medium |\n\n**Goes stale fastest:** §2 → `cloud-models-providers-and-primitives`, §13 → `cloud-migration-sovereignty-and-ai-workloads`, and the market figures throughout. **Essentially never\nstale:** §1 → `cloud-models-providers-and-primitives`, §4 → `cloud-architecture-and-resilience`, §6.2 → `cloud-architecture-and-resilience`–6.3's principles, §8 → `cloud-cost-security-and-operations`, §9 → `cloud-cost-security-and-operations`, §10 → `cloud-migration-sovereignty-and-ai-workloads`, §15.\n\n---\n\n## §18. The Canon\n\n### 18.1 Books\n\n| Author | Work | Why |\n|---|---|---|\n| **Kleppmann** | ***Designing Data-Intensive Applications*** | The forces underneath every distributed cloud decision |\n| **Beyer, Jones, Petoff & Murphy** | ***Site Reliability Engineering*** + *The SRE Workbook* | **Free online.** §9 → `cloud-cost-security-and-operations` and §6 → `cloud-architecture-and-resilience` in full |\n| **Nygard** | ***Release It!*** (2nd ed.) | ⚠️ **Still the best book on production failure modes** — §6 → `cloud-architecture-and-resilience` |\n| **Adzic & Chatley** et al. | *Serverless* literature | §5 → `cloud-architecture-and-resilience`'s trade-offs |\n| **Burns, Beda, Hightower** | *Kubernetes: Up and Running* | §5 → `cloud-architecture-and-resilience` |\n| **Morris** | ***Infrastructure as Code*** (2nd ed.) | §4 → `cloud-architecture-and-resilience` |\n| **Forsgren, Humble & Kim** | ***Accelerate*** | The delivery-performance evidence base |\n| **Storment & Fuller** | ***Cloud FinOps*** (2nd ed.) | §7 → `cloud-cost-security-and-operations`, and the standard text |\n| **Newman** | *Building Microservices* | §12 → `cloud-migration-sovereignty-and-ai-workloads`'s coupling questions |\n| **Vogels \u002F Werner Vogels' writing** | *All Things Distributed* | Design-for-failure from the source |\n\n### 18.2 Primary sources — use these over blogs\n**AWS Well-Architected Framework**, **Azure Well-Architected Framework**, and **Google\nCloud Architecture Framework** (all free, all genuinely good, all vendor-shaped).\n**Microsoft's Cloud Design Patterns** and **AWS Prescriptive Guidance**. **The provider\npost-incident reports** — ⚠️ **read these; they are the best free distributed-systems\neducation available.** **The CNCF landscape** (with a sense of humour about its size).\n**FinOps Foundation** (framework, **State of FinOps**, and **FOCUS**). **Flexera State of\nthe Cloud** (annual, 15th edition). **Synergy Research** and **Omdia** for market data.\n**EUR-Lex** for the Data Act text itself, rather than summaries of it.\n\n### 18.3 People and commentary\n**Corey Quinn** (*Last Week in AWS* — ⚠️ **the sharpest and funniest cloud-cost commentary\nanywhere, and unusually willing to say what vendors won't**), **Werner Vogels**,\n**Charity Majors** (observability, and on-call as a design constraint), **Liz Fong-Jones**,\n**Gergely Orosz** (*The Pragmatic Engineer*), **Forrest Brazeal**, **Adrian Cockcroft**,\n**Kelsey Hightower** (⚠️ **including his scepticism about Kubernetes over-adoption**),\n**Simon Wardley** (mapping, and the strategic view of commoditization).\n\n---\n\n## §19. Quick Reference\n\n### 19.1 Design checklist\n- [ ] RTO and RPO defined **per workload**, not globally (§6.3 → `cloud-architecture-and-resilience`)\n- [ ] **Dependency map including SaaS and their underlying providers** (§6.3 → `cloud-architecture-and-resilience`)\n- [ ] ⚠️ Does anything silently depend on a single region — including IAM, DNS, CI\u002FCD, or\n      monitoring? (§6.2 → `cloud-architecture-and-resilience`)\n- [ ] Degraded mode designed, not just up\u002Fdown (§6.3 → `cloud-architecture-and-resilience`)\n- [ ] Failover **tested**, not documented (§6.3 → `cloud-architecture-and-resilience`)\n- [ ] Tagging strategy enforced before spend scales (§7 → `cloud-cost-security-and-operations`)\n- [ ] Commitment and spot mix modelled (§7.2 → `cloud-cost-security-and-operations`)\n- [ ] Egress and cross-AZ paths deliberate (§11.3 → `cloud-migration-sovereignty-and-ai-workloads`)\n- [ ] Workload identity, not long-lived keys (§8 → `cloud-cost-security-and-operations`)\n- [ ] Least privilege from deny-all (§8 → `cloud-cost-security-and-operations`)\n- [ ] Logging on, retained, and outside the failure domain (§8 → `cloud-cost-security-and-operations`, §9 → `cloud-cost-security-and-operations`)\n- [ ] SLOs from user-visible behaviour; alerts on symptoms (§9 → `cloud-cost-security-and-operations`)\n- [ ] IaC for everything; no console changes in prod (§4 → `cloud-architecture-and-resilience`)\n- [ ] ⚠️ EU contracts reviewed against the **12 Jan 2027** ban (§11.2 → `cloud-migration-sovereignty-and-ai-workloads`)\n- [ ] Exit path documented — data formats, APIs, IAM, not just egress cost (§12 → `cloud-migration-sovereignty-and-ai-workloads`)\n\n### 19.2 Numbers worth remembering\n- **~29%** of IaaS\u002FPaaS spend wasted (Flexera 2026) — **up, not down**\n- **Fewer than half** of orgs use any one commitment discount\n- **Spot: 60–90% cheaper** for interruptible work\n- **GPUs ~18%** of AI-forward spend; **30–40% utilization** when statically provisioned\n- **AI ~19%** of cloud spend, from ~8% in 2023\n- **~21%** of workloads repatriated; cloud still growing\n- **AWS us-east-1 outage: ~14–15 hours**, October 2025\n- **⚠️ 12 January 2027** — EU egress and switching fee ban\n\n### 19.3 Triage\n| Symptom | First look |\n|---|---|\n| Bill jumped and nobody knows why | Untagged resources; egress; a new AI workload (§7 → `cloud-cost-security-and-operations`) |\n| Bill high but usage flat | Idle resources, orphaned storage, no commitments (§7.2 → `cloud-cost-security-and-operations`) |\n| Outage despite multi-AZ | ⚠️ Regional control-plane dependency (§6.2 → `cloud-architecture-and-resilience`) |\n| Failed over but still down | Data tier, CI\u002FCD, or monitoring in the failed region (§6.2 → `cloud-architecture-and-resilience`) |\n| Serverless costs more than expected | Sustained volume — the curve inverted (§5 → `cloud-architecture-and-resilience`) |\n| Kubernetes consuming the team | You may not need it (§5 → `cloud-architecture-and-resilience`) |\n| Cross-AZ charges surprising | Chatty services spanning zones (§11.3 → `cloud-migration-sovereignty-and-ai-workloads`) |\n| \"We're multi-cloud\" but can't fail over | You have workload distribution, not portability (§12 → `cloud-migration-sovereignty-and-ai-workloads`) |\n| Compliance auditor unsatisfied despite EU region | ⚠️ Residency ≠ sovereignty (§11.1 → `cloud-migration-sovereignty-and-ai-workloads`) |\n\n---\n\n## §20. Sources and Method\n\n**Method.** Narrative review. The architectural material — §1 → `cloud-models-providers-and-primitives`, §4 → `cloud-architecture-and-resilience`, §6.3 → `cloud-architecture-and-resilience`'s principles, §8 → `cloud-cost-security-and-operations`,\n§9 → `cloud-cost-security-and-operations`, §10 → `cloud-migration-sovereignty-and-ai-workloads`, §15 — rests on the provider Well-Architected frameworks, SRE literature, and\nlong-stable practice, and does not move much. **The economics, market structure, and\nregulation move fast**, and those were verified in **August 2026** with four targeted\nsearches; every such claim is flagged **[VERSIONED]** and carries a decay rating in §17.\n\n**Search log** (August 2026): cloud market share and AI capex · EU Data Act cloud\nswitching, egress fees, and sovereignty · the October–November 2025 outages and resilience\nlessons · FinOps, cloud waste, and repatriation.\n\n**Primary and near-primary sources consulted (selected):**\n- **Market**: Synergy Research figures as reported via Statista and multiple trackers;\n  Omdia's Q4 2025 cloud infrastructure analysis; company earnings coverage (Amazon,\n  Microsoft, Alphabet Q1 2026). **⚠️ The methodological explanation for the divergent share\n  figures came from a tracker that showed its working** — I've reproduced the reasoning\n  rather than picking a number\n- **EU Data Act**: legal analyses from **Alston & Bird**, **McCann FitzGerald**,\n  **Kemp IT Law**, **CMS**, and **Turing Law** for the switching-charge provisions,\n  transition rules, and enforcement structure; **CSIS** for the Data Act \u002F DMA friction\n  analysis; contemporaneous coverage for the June 2026 Cloud and AI Development Act\n  proposal\n- **Outages**: incident analyses from **Cockroach Labs' 2025 outage review**,\n  **DevOps.com**, **Datacenter Knowledge**, **INE**, and independent write-ups; AWS's own\n  post-mortem as reported. **⚠️ I did not read the provider post-incident reports directly**\n- **FinOps**: **Flexera *State of the Cloud* 2026** (n=753) via its own materials and\n  secondary reporting; **FinOps Foundation *State of FinOps* 2026**; CIO Dive's coverage of\n  the repatriation and waste figures\n\n**Confidence statement.** **High confidence** in §1 → `cloud-models-providers-and-primitives`, §3–§5 → `cloud-models-providers-and-primitives`, `cloud-architecture-and-resilience`, §6.2 → `cloud-architecture-and-resilience`–6.3, §8–§10 → `cloud-cost-security-and-operations`, `cloud-migration-sovereignty-and-ai-workloads`, §15 — these\nare architectural principles and widely-corroborated practice. **High confidence in the\nEU Data Act dates and structure** in §11.2 → `cloud-migration-sovereignty-and-ai-workloads`, which came from multiple independent legal\nanalyses agreeing on the 12 September 2025 and **12 January 2027** dates; the Regulation\nnumber and text are checkable at EUR-Lex and I'd recommend doing so before relying on it\ncommercially. **High confidence in the outage facts** in §6.1 → `cloud-architecture-and-resilience`, which are consistently\nreported across many independent sources, though **root-cause descriptions are\nsummarized from secondary analyses rather than from the provider post-incident reports\nthemselves.**\n\n**⚠️ Lower confidence, deliberately flagged, on the numbers.** **Market share figures\ngenuinely conflict by 4+ percentage points across reputable sources for methodological\nreasons I've explained in §2 → `cloud-models-providers-and-primitives`** — treat any single figure as directional. **Capex figures\nare company guidance, not actuals.** **Waste, repatriation, and GPU-utilization figures\nare survey-based** (Flexera n=753, FinOps Foundation n≈523 for some measures), carry\nself-report bias, and vary between surveys — Harness reported ~21% waste against Flexera's\n29% in overlapping periods. **The us-east-1 concentration figures (~69% of usage, 44%+ of\nrequests) come from different methodologies measuring different things** and should not be\ncombined. Several FinOps-adjacent sources are **vendor-published with an obvious interest\nin the size of the problem they sell tooling for**, which I've weighted accordingly. §16 is\nopinion labelled as such.\n","# Cloud Computing: Anti-Patterns, Contested Questions, Currency, and Canon\n\n> **Part 5 of 5** of the *Cloud Computing* reference (plugin `cloud-computing`), covering §15–§20. Sibling skills: `cloud-models-providers-and-primitives` (§0–§3), `cloud-architecture-and-resilience` (§4–§6), `cloud-cost-security-and-operations` (§7–§9), `cloud-migration-sovereignty-and-ai-workloads` (§10–§14). Section numbers are shared across the set; a reference written as §N → `skill` points into that sibling skill.\n>\n> **Currency:** Verified August 2026. See §17 below for the currency snapshot and what goes stale first.\n\n> **How to read this.** Reference, not a certification guide. Three markers:\n> - **[DURABLE]** — architectural principles and trade-offs that survive provider churn.\n> - **[VERSIONED]** — market data, pricing, service names, regulation. **Verify before\n>   quoting.**\n> - **[CONTESTED]** — genuine disagreement.\n>\n> **⚠️ GOTCHA** boxes mark the mistakes that produce outages or surprise invoices.\n>\n> **Provider-neutral by default**, with AWS\u002FAzure\u002FGCP names where the concept needs an\n> anchor. Service names change; the primitives don't.\n>\n> **The three framings that organize everything below:**\n> 1. **The cloud is someone else's computer, rented, with an API and a metered bill.**\n>    Every genuine advantage (elasticity, global reach, managed services, capex→opex) and\n>    every genuine failure mode (cost surprise, lock-in, shared fate, opaque outages)\n>    follows from that one sentence.\n> 2. **⚠️ Your failure domain is always larger than your architecture diagram says.**\n>    The 2025 outages taught this expensively: multi-AZ deployments failed because of\n>    control-plane dependencies nobody had drawn (§6 → `cloud-architecture-and-resilience`). **Map what you actually depend on,\n>    not what you think you depend on.**\n> 3. **Cost is an architectural property, not an operational afterthought.** It is\n>    determined by design decisions — data placement, service selection, egress paths —\n>    and **discovered on an invoice 30 days later** (§7 → `cloud-cost-security-and-operations`). Design for it up front or pay\n>    for it forever.\n\n---\n\n## §15. Anti-Patterns\n\n| Anti-pattern | Why |\n|---|---|\n| Lift-and-shift everything, then be surprised by the bill | Fixed-capacity design on per-hour billing (§10 → `cloud-migration-sovereignty-and-ai-workloads`) |\n| Treating cost as an ops problem | ⚠️ **It's set at design time** (§7 → `cloud-cost-security-and-operations`) |\n| No tagging strategy | **You can't allocate what you can't attribute** (§7 → `cloud-cost-security-and-operations`) |\n| Ignoring commitment discounts | **Fewer than half of orgs use any** — free money (§7.1 → `cloud-cost-security-and-operations`) |\n| Never using spot for interruptible work | 60–90% cheaper (§3 → `cloud-models-providers-and-primitives`, §13 → `cloud-migration-sovereignty-and-ai-workloads`) |\n| Provisioned block storage sized for peak, forever | You pay for the ask, not the use (§3 → `cloud-models-providers-and-primitives`) |\n| No storage lifecycle policy | Minutes to set, runs forever (§3 → `cloud-models-providers-and-primitives`) |\n| **Assuming multi-AZ protects against regional failure** | ⚠️ **The single most expensive lesson of 2025** (§6 → `cloud-architecture-and-resilience`) |\n| Multi-region compute with single-region data | The dependency that actually took people down (§6.2 → `cloud-architecture-and-resilience`) |\n| Not auditing whether failover routes through us-east-1 | Global services anchor there (§6.2 → `cloud-architecture-and-resilience`) |\n| Monitoring and alerting behind the same provider you're monitoring | ⚠️ **Failed exactly this way in 2025** (§6.2 → `cloud-architecture-and-resilience`) |\n| CI\u002FCD and IaC state in the region you'd fail over *from* | Can't deploy the fix (§6.3 → `cloud-architecture-and-resilience`) |\n| Untested failover runbook | It's a hypothesis, not a plan (§6.3 → `cloud-architecture-and-resilience`) |\n| Binary up\u002Fdown with no degraded mode | Read-only beats down, and costs far less (§6.3 → `cloud-architecture-and-resilience`) |\n| Long-lived access keys | ⚠️ **The most common breach vector.** Use workload identity (§8 → `cloud-cost-security-and-operations`) |\n| Wildcard IAM policies that were \"temporary\" | They never are (§8 → `cloud-cost-security-and-operations`) |\n| Treating a VPC as a trust boundary | Network location ≠ authorization (§8 → `cloud-cost-security-and-operations`) |\n| Assuming provider compliance certification = your compliance | It enables it; config is yours (§8 → `cloud-cost-security-and-operations`) |\n| Alerting on CPU instead of user-visible symptoms | Alert fatigue is a reliability problem (§9 → `cloud-cost-security-and-operations`) |\n| **Assuming data residency equals sovereignty** | ⚠️ **The \"residency illusion\"** — CLOUD Act reach (§11.1 → `cloud-migration-sovereignty-and-ai-workloads`) |\n| Signing or auto-renewing EU cloud contracts without the 2027 ban in view | ⚠️ **Locks in an outdated fee structure** (§11.2 → `cloud-migration-sovereignty-and-ai-workloads`) |\n| Thinking egress fees were the whole of lock-in | APIs, formats and IAM outlast the invoice line (§11.2 → `cloud-migration-sovereignty-and-ai-workloads`, §12 → `cloud-migration-sovereignty-and-ai-workloads`) |\n| Chatty service mesh spanning AZs | Cross-AZ transfer is a real, invisible cost (§11.3 → `cloud-migration-sovereignty-and-ai-workloads`) |\n| Kubernetes for three services and five engineers | ⚠️ **Adopting a platform team's problem without the platform team** (§5 → `cloud-architecture-and-resilience`) |\n| Serverless for sustained high-volume load | The cost curve inverts (§5 → `cloud-architecture-and-resilience`) |\n| Multi-cloud portability as a default goal | Lowest common denominator, multiplied ops (§12 → `cloud-migration-sovereignty-and-ai-workloads`) |\n| Accidental lock-in | ⚠️ **You paid the abstraction cost and got locked in anyway** (§12 → `cloud-migration-sovereignty-and-ai-workloads`) |\n| Statically provisioned GPU fleets | **30–40% utilization; fastest-growing waste** (§13 → `cloud-migration-sovereignty-and-ai-workloads`) |\n| Training on on-demand instead of spot with checkpointing | Bursty and interruption-tolerant by nature (§13 → `cloud-migration-sovereignty-and-ai-workloads`) |\n| Retrofitting AI cost attribution later | **What the 2026 waste reversal is measuring** (§7.1 → `cloud-cost-security-and-operations`, §13 → `cloud-migration-sovereignty-and-ai-workloads`) |\n| Quoting a cloud market share figure without its source | ⚠️ **They differ by 4+ points on methodology** (§2 → `cloud-models-providers-and-primitives`) |\n\n---\n\n## §16. Contested Questions\n\n**16.1 Is cloud cheaper than on-prem?** *For*: no capex, elastic, no datacenter staff,\nyou pay for what you use. *Against*: at **steady, predictable, high utilization**, owned\nhardware is frequently cheaper — and that's exactly the profile that gets repatriated.\n**[CONTESTED, and the honest answer is workload-dependent.]** The 2026 data supports a\nmiddle reading: **~21% of workloads repatriated, but new cloud workloads outstrip the\nexits.** ⚠️ **From 2027 the EU removes exit-cost distortion from this comparison, which\nmeans the TCO math should genuinely be re-run** rather than assumed.\n\n**16.2 Is repatriation a real trend or vendor noise?** Real but bounded. **Selective\nrepatriation of steady-state, high-utilization workloads is happening**, driven partly by\nFinOps tooling finally making the comparison visible. **Wholesale cloud exit is not.**\n\n**16.3 Multi-cloud: insurance or expensive complexity?** §12 → `cloud-migration-sovereignty-and-ai-workloads`. The 2025 outages strengthened\nthe resilience argument; the operational reality strengthened the complexity argument.\n**Both got more true at once.**\n\n**16.4 Is Kubernetes over-adopted?** **[CONTESTED but leaning yes.]** Excellent at scale\nwith a platform team, poor fit below that, and adopted for reasons that are often about\nrésumés and perceived sophistication rather than requirements.\n\n**16.5 Does serverless deliver?** *For*: genuinely excellent for event-driven, spiky, and\nlow-volume workloads; operational burden near zero. *Against*: cost inversion at sustained\nvolume, cold starts, testing friction, vendor coupling. **The pendulum settled around\n\"right tool for specific shapes\" rather than a default.**\n\n**16.6 Is sovereign cloud meaningful or theatre?** *For*: the CLOUD Act tension is real\nand residency genuinely doesn't solve it. *Against*: much of what's marketed as sovereign\nis residency with extra branding. **⚠️ The EU's proposed Cloud and AI Development Act\n(June 2026) is an attempt to settle this by defining graded sovereignty levels** — which\nis itself an admission that the term is currently unregulated marketing.\n\n**16.7 Are the hyperscalers over-building for AI?** ⚠️ **Genuinely unknown, and the stakes\nare large.** ~$570B+ of combined 2026 capex guidance is a bet that AI demand\nmaterializes. Revenue growth currently supports it. **If it disappoints, the write-downs\nwould be historically large** — and buyers should factor that uncertainty into\nlong-horizon commitments.\n\n---\n\n## §17. Currency Snapshot — verified August 2026\n\n| Thing | Status as of Aug 2026 | Decay risk |\n|---|---|---|\n| **Market structure** | Q1 2026 cloud infrastructure spend **~$129B, +35% YoY**, ninth consecutive quarter of accelerating YoY growth. Big Three **~60–68%** depending on methodology. ⚠️ **Synergy: AWS 28% \u002F Azure 21% \u002F GCP 14%. Other sources: 30\u002F24–25\u002F13 or 32\u002F23\u002F11** — **cite source and scope** (§2 → `cloud-models-providers-and-primitives`). Azure and GCP growing faster off smaller bases; **Microsoft doesn't break out Azure revenue, only growth rate** | **High** |\n| **Capex** | 2026 guidance: **Amazon ~$200B, Microsoft ~$190B, Alphabet ~$180–190B.** Guidance, not actuals | **High** |\n| **AI share of cloud spend** | **~19% in 2026, up from ~8% in 2023** | **High** |\n| **⚠️ Cloud waste** | **Flexera 2026: 29% of IaaS\u002FPaaS wasted, up from 27% — reversing a five-year decline**, attributed to **AI adoption outpacing tagging\u002Fattribution\u002Fgovernance**. n=753. **Managing cloud spend now the #1 priority (84%), above security** | Medium |\n| **FinOps adoption** | **FinOps teams ~63%, CCoE ~71%.** ⚠️ **Fewer than half use any one commitment discount per provider.** **AI cost management the top future priority (98%)** and most-wanted skillset. **FOCUS** standardizing cross-provider cost data. Practitioners report **the easy waste is gone** — remaining savings are smaller and harder | Medium |\n| **Repatriation** | **~21% of workloads repatriated**, but **new cloud workloads outstrip exits — cloud continues to grow** | Medium |\n| **GPU economics** | **GPUs ~18% of spend at AI-forward orgs (up from ~4% in 2023); statically provisioned fleets at 30–40% utilization.** Idle accelerators = fastest-growing waste category | **High** |\n| **⚠️ Oct–Nov 2025 outages** | **AWS us-east-1, 20 Oct, ~14–15h** — latent **race condition in DynamoDB's DNS management** wiped DNS records, cascading across dozens of services. **Azure, 29 Oct** — **Azure Front Door** \u002F global identity. **Cloudflare, 18 Nov** — Bot Management config **doubled a feature file past a hard-coded limit**, crashing proxies globally; **engineers spent ~2h believing it was a DDoS**. Common pattern: **subtle single-subsystem defect → global cascade**; diagnosis is **immature blast-radius modelling**, not human error | Low (historical) |\n| **us-east-1 concentration** | Reported at **~69% of AWS usage** by one analysis and **44%+ of AWS requests** by another. **Global operations (parts of IAM, Route 53) anchor there** | Medium |\n| **⚠️ EU Data Act** | Reg. (EU) 2023\u002F2854. In force 11 Jan 2024. **Switching rights enforceable 12 Sept 2025.** **12 Sept 2026: products released after must have user data accessible by default.** **⚠️ 12 January 2027: all switching charges including egress fees banned outright** — IaaS, PaaS and SaaS, **for every provider serving EU customers**. Transition allows only direct, transparent, pre-agreed cost. Enforcement by **national authorities**, penalties set in **local legislation**. AWS\u002FAzure\u002FGoogle waived full-exit egress in 2024 (narrower than 2027 requires) | **High** |\n| **DMA \u002F sovereignty** | **DMA investigating AWS and Azure as potential gatekeepers**; ⚠️ **analysts flag genuine friction with the Data Act** — conflicting timelines, split enforcement (member states vs. Commission). **3 June 2026: Commission proposed the Cloud and AI Development Act** (Tech Sovereignty Package) to define **graded sovereignty levels** — ⚠️ **not yet passed; details will change** | **High** |\n| **Sustainability** | Well-Architected pillar; **>half of orgs have or plan a sustainability initiative including cloud carbon tracking** | Medium |\n\n**Goes stale fastest:** §2 → `cloud-models-providers-and-primitives`, §13 → `cloud-migration-sovereignty-and-ai-workloads`, and the market figures throughout. **Essentially never\nstale:** §1 → `cloud-models-providers-and-primitives`, §4 → `cloud-architecture-and-resilience`, §6.2 → `cloud-architecture-and-resilience`–6.3's principles, §8 → `cloud-cost-security-and-operations`, §9 → `cloud-cost-security-and-operations`, §10 → `cloud-migration-sovereignty-and-ai-workloads`, §15.\n\n---\n\n## §18. The Canon\n\n### 18.1 Books\n\n| Author | Work | Why |\n|---|---|---|\n| **Kleppmann** | ***Designing Data-Intensive Applications*** | The forces underneath every distributed cloud decision |\n| **Beyer, Jones, Petoff & Murphy** | ***Site Reliability Engineering*** + *The SRE Workbook* | **Free online.** §9 → `cloud-cost-security-and-operations` and §6 → `cloud-architecture-and-resilience` in full |\n| **Nygard** | ***Release It!*** (2nd ed.) | ⚠️ **Still the best book on production failure modes** — §6 → `cloud-architecture-and-resilience` |\n| **Adzic & Chatley** et al. | *Serverless* literature | §5 → `cloud-architecture-and-resilience`'s trade-offs |\n| **Burns, Beda, Hightower** | *Kubernetes: Up and Running* | §5 → `cloud-architecture-and-resilience` |\n| **Morris** | ***Infrastructure as Code*** (2nd ed.) | §4 → `cloud-architecture-and-resilience` |\n| **Forsgren, Humble & Kim** | ***Accelerate*** | The delivery-performance evidence base |\n| **Storment & Fuller** | ***Cloud FinOps*** (2nd ed.) | §7 → `cloud-cost-security-and-operations`, and the standard text |\n| **Newman** | *Building Microservices* | §12 → `cloud-migration-sovereignty-and-ai-workloads`'s coupling questions |\n| **Vogels \u002F Werner Vogels' writing** | *All Things Distributed* | Design-for-failure from the source |\n\n### 18.2 Primary sources — use these over blogs\n**AWS Well-Architected Framework**, **Azure Well-Architected Framework**, and **Google\nCloud Architecture Framework** (all free, all genuinely good, all vendor-shaped).\n**Microsoft's Cloud Design Patterns** and **AWS Prescriptive Guidance**. **The provider\npost-incident reports** — ⚠️ **read these; they are the best free distributed-systems\neducation available.** **The CNCF landscape** (with a sense of humour about its size).\n**FinOps Foundation** (framework, **State of FinOps**, and **FOCUS**). **Flexera State of\nthe Cloud** (annual, 15th edition). **Synergy Research** and **Omdia** for market data.\n**EUR-Lex** for the Data Act text itself, rather than summaries of it.\n\n### 18.3 People and commentary\n**Corey Quinn** (*Last Week in AWS* — ⚠️ **the sharpest and funniest cloud-cost commentary\nanywhere, and unusually willing to say what vendors won't**), **Werner Vogels**,\n**Charity Majors** (observability, and on-call as a design constraint), **Liz Fong-Jones**,\n**Gergely Orosz** (*The Pragmatic Engineer*), **Forrest Brazeal**, **Adrian Cockcroft**,\n**Kelsey Hightower** (⚠️ **including his scepticism about Kubernetes over-adoption**),\n**Simon Wardley** (mapping, and the strategic view of commoditization).\n\n---\n\n## §19. Quick Reference\n\n### 19.1 Design checklist\n- [ ] RTO and RPO defined **per workload**, not globally (§6.3 → `cloud-architecture-and-resilience`)\n- [ ] **Dependency map including SaaS and their underlying providers** (§6.3 → `cloud-architecture-and-resilience`)\n- [ ] ⚠️ Does anything silently depend on a single region — including IAM, DNS, CI\u002FCD, or\n      monitoring? (§6.2 → `cloud-architecture-and-resilience`)\n- [ ] Degraded mode designed, not just up\u002Fdown (§6.3 → `cloud-architecture-and-resilience`)\n- [ ] Failover **tested**, not documented (§6.3 → `cloud-architecture-and-resilience`)\n- [ ] Tagging strategy enforced before spend scales (§7 → `cloud-cost-security-and-operations`)\n- [ ] Commitment and spot mix modelled (§7.2 → `cloud-cost-security-and-operations`)\n- [ ] Egress and cross-AZ paths deliberate (§11.3 → `cloud-migration-sovereignty-and-ai-workloads`)\n- [ ] Workload identity, not long-lived keys (§8 → `cloud-cost-security-and-operations`)\n- [ ] Least privilege from deny-all (§8 → `cloud-cost-security-and-operations`)\n- [ ] Logging on, retained, and outside the failure domain (§8 → `cloud-cost-security-and-operations`, §9 → `cloud-cost-security-and-operations`)\n- [ ] SLOs from user-visible behaviour; alerts on symptoms (§9 → `cloud-cost-security-and-operations`)\n- [ ] IaC for everything; no console changes in prod (§4 → `cloud-architecture-and-resilience`)\n- [ ] ⚠️ EU contracts reviewed against the **12 Jan 2027** ban (§11.2 → `cloud-migration-sovereignty-and-ai-workloads`)\n- [ ] Exit path documented — data formats, APIs, IAM, not just egress cost (§12 → `cloud-migration-sovereignty-and-ai-workloads`)\n\n### 19.2 Numbers worth remembering\n- **~29%** of IaaS\u002FPaaS spend wasted (Flexera 2026) — **up, not down**\n- **Fewer than half** of orgs use any one commitment discount\n- **Spot: 60–90% cheaper** for interruptible work\n- **GPUs ~18%** of AI-forward spend; **30–40% utilization** when statically provisioned\n- **AI ~19%** of cloud spend, from ~8% in 2023\n- **~21%** of workloads repatriated; cloud still growing\n- **AWS us-east-1 outage: ~14–15 hours**, October 2025\n- **⚠️ 12 January 2027** — EU egress and switching fee ban\n\n### 19.3 Triage\n| Symptom | First look |\n|---|---|\n| Bill jumped and nobody knows why | Untagged resources; egress; a new AI workload (§7 → `cloud-cost-security-and-operations`) |\n| Bill high but usage flat | Idle resources, orphaned storage, no commitments (§7.2 → `cloud-cost-security-and-operations`) |\n| Outage despite multi-AZ | ⚠️ Regional control-plane dependency (§6.2 → `cloud-architecture-and-resilience`) |\n| Failed over but still down | Data tier, CI\u002FCD, or monitoring in the failed region (§6.2 → `cloud-architecture-and-resilience`) |\n| Serverless costs more than expected | Sustained volume — the curve inverted (§5 → `cloud-architecture-and-resilience`) |\n| Kubernetes consuming the team | You may not need it (§5 → `cloud-architecture-and-resilience`) |\n| Cross-AZ charges surprising | Chatty services spanning zones (§11.3 → `cloud-migration-sovereignty-and-ai-workloads`) |\n| \"We're multi-cloud\" but can't fail over | You have workload distribution, not portability (§12 → `cloud-migration-sovereignty-and-ai-workloads`) |\n| Compliance auditor unsatisfied despite EU region | ⚠️ Residency ≠ sovereignty (§11.1 → `cloud-migration-sovereignty-and-ai-workloads`) |\n\n---\n\n## §20. Sources and Method\n\n**Method.** Narrative review. The architectural material — §1 → `cloud-models-providers-and-primitives`, §4 → `cloud-architecture-and-resilience`, §6.3 → `cloud-architecture-and-resilience`'s principles, §8 → `cloud-cost-security-and-operations`,\n§9 → `cloud-cost-security-and-operations`, §10 → `cloud-migration-sovereignty-and-ai-workloads`, §15 — rests on the provider Well-Architected frameworks, SRE literature, and\nlong-stable practice, and does not move much. **The economics, market structure, and\nregulation move fast**, and those were verified in **August 2026** with four targeted\nsearches; every such claim is flagged **[VERSIONED]** and carries a decay rating in §17.\n\n**Search log** (August 2026): cloud market share and AI capex · EU Data Act cloud\nswitching, egress fees, and sovereignty · the October–November 2025 outages and resilience\nlessons · FinOps, cloud waste, and repatriation.\n\n**Primary and near-primary sources consulted (selected):**\n- **Market**: Synergy Research figures as reported via Statista and multiple trackers;\n  Omdia's Q4 2025 cloud infrastructure analysis; company earnings coverage (Amazon,\n  Microsoft, Alphabet Q1 2026). **⚠️ The methodological explanation for the divergent share\n  figures came from a tracker that showed its working** — I've reproduced the reasoning\n  rather than picking a number\n- **EU Data Act**: legal analyses from **Alston & Bird**, **McCann FitzGerald**,\n  **Kemp IT Law**, **CMS**, and **Turing Law** for the switching-charge provisions,\n  transition rules, and enforcement structure; **CSIS** for the Data Act \u002F DMA friction\n  analysis; contemporaneous coverage for the June 2026 Cloud and AI Development Act\n  proposal\n- **Outages**: incident analyses from **Cockroach Labs' 2025 outage review**,\n  **DevOps.com**, **Datacenter Knowledge**, **INE**, and independent write-ups; AWS's own\n  post-mortem as reported. **⚠️ I did not read the provider post-incident reports directly**\n- **FinOps**: **Flexera *State of the Cloud* 2026** (n=753) via its own materials and\n  secondary reporting; **FinOps Foundation *State of FinOps* 2026**; CIO Dive's coverage of\n  the repatriation and waste figures\n\n**Confidence statement.** **High confidence** in §1 → `cloud-models-providers-and-primitives`, §3–§5 → `cloud-models-providers-and-primitives`, `cloud-architecture-and-resilience`, §6.2 → `cloud-architecture-and-resilience`–6.3, §8–§10 → `cloud-cost-security-and-operations`, `cloud-migration-sovereignty-and-ai-workloads`, §15 — these\nare architectural principles and widely-corroborated practice. **High confidence in the\nEU Data Act dates and structure** in §11.2 → `cloud-migration-sovereignty-and-ai-workloads`, which came from multiple independent legal\nanalyses agreeing on the 12 September 2025 and **12 January 2027** dates; the Regulation\nnumber and text are checkable at EUR-Lex and I'd recommend doing so before relying on it\ncommercially. **High confidence in the outage facts** in §6.1 → `cloud-architecture-and-resilience`, which are consistently\nreported across many independent sources, though **root-cause descriptions are\nsummarized from secondary analyses rather than from the provider post-incident reports\nthemselves.**\n\n**⚠️ Lower confidence, deliberately flagged, on the numbers.** **Market share figures\ngenuinely conflict by 4+ percentage points across reputable sources for methodological\nreasons I've explained in §2 → `cloud-models-providers-and-primitives`** — treat any single figure as directional. **Capex figures\nare company guidance, not actuals.** **Waste, repatriation, and GPU-utilization figures\nare survey-based** (Flexera n=753, FinOps Foundation n≈523 for some measures), carry\nself-report bias, and vary between surveys — Harness reported ~21% waste against Flexera's\n29% in overlapping periods. **The us-east-1 concentration figures (~69% of usage, 44%+ of\nrequests) come from different methodologies measuring different things** and should not be\ncombined. Several FinOps-adjacent sources are **vendor-published with an obvious interest\nin the size of the problem they sell tooling for**, which I've weighted accordingly. §16 is\nopinion labelled as such.\n",{"name":5,"description":7},"develop","ae32736a1aabddabca5d2c96d665aa0ebc7d8700","src\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fcloud-computing\u002Fskills\u002Fcloud-reference",3,"https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey","https:\u002F\u002Fthe-vibey-project.github.io\u002Fvibey\u002F","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project","https:\u002F\u002Fvibewithadam.matthewsteinberger.com","freedom for software engineers","2026-11-15 05:47:42",{"source":8,"skillName":5,"commands":35,"prompt":45,"zipAvailable":46},[36,39,42],{"tool":37,"command":38},"npx","npx skills add the-vibey-project\u002Fvibey --skill cloud-reference",{"tool":40,"command":41},"bunx","bunx skills add the-vibey-project\u002Fvibey --skill cloud-reference",{"tool":43,"command":44},"pnpm","pnpm dlx skills add the-vibey-project\u002Fvibey --skill cloud-reference","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fcloud-computing\u002Fskills\u002Fcloud-reference\u002FSKILL.md",true,[],[49,56,63,70,77,84],{"skillNo":50,"name":51,"nameZh":6,"description":52,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":53,"categories":54,"keywords":55,"publishedAt":19},"01M7AT1FQJAE1FQYAMDTF490Y4","white-paper-writing","Use when writing white papers, technical reports, policy documents, thought leadership pieces, or any long-form authoritative document. Triggers on requests to write, structure, outline, or review white papers, technical reports, position papers, policy briefs, or evidence-driven arguments. Also use for B2B content marketing strategy, technical documentation requiring authority, academic-style reports, or any document that must educate while persuading. Use when someone asks \"how should I structure this\" for a long-form document.","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fwriting-craft\u002Fskills\u002Fwhite-paper-writing\u002FSKILL.md",[],[],{"skillNo":57,"name":58,"nameZh":6,"description":59,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":60,"categories":61,"keywords":62,"publishedAt":19},"01M7AT16BJRJJKK2MASXRZS76G","prose-technical-mechanics","Use when editing or writing prose at a technical\u002Fmechanical level: choosing between Anglo-Saxon and Latinate diction, placing information in topic vs stress positions, writing periodic vs cumulative sentences, handling sentence rhythm, punctuation as logic, transitions, paragraph structure, or phonesthetics. Use when someone wants to understand WHY prose works or doesn't, not just surface-level style advice. Triggers on word choice, sentence structure, emphasis, rhythm, diction, or any deep writing mechanics question.","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fwriting-craft\u002Fskills\u002Fprose-technical-mechanics\u002FSKILL.md",[],[],{"skillNo":64,"name":65,"nameZh":6,"description":66,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":67,"categories":68,"keywords":69,"publishedAt":19},"01M7AT0W9HT0ZZ9Y287NW99MR3","prose-mechanics","Use when writing, editing, or analyzing prose style; when trying to avoid generic AI-sounding writing; when applying specific prose styles such as journalistic, academic, narrative, KJV biblical, Hemingway plain, Latinate\u002FJamesian, Orwellian, Augustan, Victorian, Early Modern (Bacon, Browne, Milton), or free indirect discourse; or when coaching someone on writing voice and mechanics. Triggers on any request to improve, analyze, or write in a specific prose style, to fix \"AI-sounding\" writing, to make prose more vivid or concrete, to vary sentence rhythm, or to understand how a particular writer achieves their effects.","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fwriting-craft\u002Fskills\u002Fprose-mechanics\u002FSKILL.md",[],[],{"skillNo":71,"name":72,"nameZh":6,"description":73,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":74,"categories":75,"keywords":76,"publishedAt":19},"01M7AT0KR635JX66X2X6F3E2CP","narrative-structure","Use when structuring stories, pitches, presentations, white papers, or any persuasive narrative; analyzing the hero's journey in fiction or non-fiction; applying StoryBrand for business communication; designing narrative arcs for thought leadership; or understanding Vogler, Snyder, Propp, or Campbell frameworks. Triggers on any narrative structure, storytelling, pitch structure, StoryBrand, hero's journey, three-act structure, Freytag, Save the Cat, narrative arc, or \"how should I structure this story\u002Fpitch\u002Fpresentation\" request.","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fwriting-craft\u002Fskills\u002Fnarrative-structure\u002FSKILL.md",[],[],{"skillNo":78,"name":79,"nameZh":6,"description":80,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":81,"categories":82,"keywords":83,"publishedAt":19},"01M7AT0AKGW09NKKZDYBT8VRDE","legal-brief-writing","Use when writing legal briefs, motions, or any court filing; when analyzing judicial cognition and persuasion; when applying elite brief-writing techniques from Roberts, Clement, Kagan, Garner; or when structuring legal arguments for maximum persuasion. Also use for autistic\u002Fneurodivergent advocates who need to understand how their natural communication style maps to elite legal advocacy. Triggers on any legal writing, brief drafting, motion preparation, or legal argument structuring request.","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fwriting-craft\u002Fskills\u002Flegal-brief-writing\u002FSKILL.md",[],[],{"skillNo":85,"name":86,"nameZh":6,"description":87,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":88,"categories":89,"keywords":90,"publishedAt":19},"01M7AT00VWPS4Q8B5VEMXX9NQT","journal-publishability-criteria","Use whenever a research paper, article or preprint is being written, revised, reviewed or published, and specifically before every publication of docs\u002Fpaper.md. Covers what legitimate journals actually gate on, in order of gating — the editor's desk screen (scope, novelty, package problems), what reviewers judge (methodological soundness, claims the evidence can carry, statistical and qualitative reporting), EQUATOR reporting guidelines, the ICMJE authorship criteria, COPE's view of text recycling, the near-universal generative-AI policy (no AI author, disclosure, author responsibility), TOP 2025 transparency standards, how to verify a journal is legitimate, submitting without an institution, and a pre-submission checklist. Triggers on requests to evaluate, review, prepare or submit a paper, to judge publishability, to write a declarations or data-availability section, to disclose AI use, or to check a journal.","https:\u002F\u002Fgithub.com\u002Fthe-vibey-project\u002Fvibey\u002Fblob\u002Fdevelop\u002Fsrc\u002Fvibey_tools\u002Fskills\u002Fplugins\u002Fwriting-craft\u002Fskills\u002Fjournal-publishability-criteria\u002FSKILL.md",[],[]]