Governing Azure by Contract: A Layer-by-Layer Guide to a Governed Azure Platform
A free field guide to governing Azure layer by layer: tenant, management groups, dev/test/prod subscriptions, roles and PIM, policy, network, security, visibility and cost.
TL;DR: Many Azure estates are not designed as one platform. They grow piece by piece: a subscription here, an Owner role there, a public endpoint nobody remembers. Governing Azure by Contract is a free field guide that does it the other way around. It governs the platform first, layer by layer, and writes each guarantee as a contract the platform enforces. It is Book 1 of the GDCD on Azure series.
When subscriptions are created outside an agreed process, resources spread faster than anyone can explain them. Ownership, cost and exposure become hard to answer. Nobody chooses that outcome. It is what happens when building starts before governing.
The guide applies Governance-Driven Contract Design (GDCD) to the cloud platform. The method always moves in the same order. Start from the business outcome. Turn it into guarantees, the promises the platform must keep. Spread the guarantees across quality, security, visibility and cost. Then write each one as a contract at the door where it can be enforced.

What is in the book
- The tenant and the management group tree. The Microsoft Entra tenant holds identity, and the tenant root management group is the highest scope for Azure Policy and Azure RBAC inheritance. The tree (Platform, Landing zones, Sandbox, Decommissioned) is where policy and access are inherited, so it is designed once and kept shallow.
- Dev, test and prod subscriptions. A separate subscription per workload environment by default, created by a subscription vending process instead of by hand, with owner, budget and tags from day one. Exceptions are documented.
- Roles: who may do what, and for how long. Least privilege with Azure RBAC, groups instead of people, and Privileged Identity Management for eligible, time-bound privileged access.
- Policy is the contract that runs. One policy baseline in dev, test and prod, so the rules tested in dev and test are the rules that apply in prod. Audit first, remediate, then deny. Only the Sandbox is looser, and every exception is a recorded exemption with an end date.
- Network, security, visibility and cost. Private by default, a security posture you can watch, central logs and alerts, and cost that can be allocated to an owner.
- Deploy by contract and the operating model. Infrastructure as code as the standard deployment path, backed by restricted write permissions and a documented break-glass process, and a 30-60-90 day rollout with clear ownership.
- Templates. An appendix with contract templates a platform team can copy.
Why now
AI workloads are arriving on Azure faster than most platforms were governed. An AI deployment is only as governed as the subscription, identity and network it runs in. Governing the platform first gives every AI workload its foundational controls before the first model is deployed.
That is why this book comes first. Its companion, AI Foundry by Contract, builds Microsoft Foundry on top of the landing zone described here.
How it was checked
Product names, roles and policy effects were checked against Microsoft documentation in November 2026. The text then went through four rounds of AI-assisted review, with several AI models prompted to read as an Azure architect, a security and FinOps buyer and a technical editor. These were review tools, not an independent audit. Each factual finding they raised was checked against Microsoft documentation before a change was accepted, and findings that did not hold up were rejected.
Read it
The guide is free to read online or download as a PDF. More interactive Azure architecture diagrams live on the Azure hub.
A useful first question for any Azure estate: could someone explain every subscription, every Owner role and every public endpoint you have today? If not, govern before you build. Trust is earned at every layer.