Practice Exams:

Microsoft AZ-104: Azure Landing Zones

Cloud & Architecture

A landing zone is the operating foundation for Azure workloads, not a subscription template with a checklist of resources. It connects governance, network access, identity, security, and the process of giving application teams usable environments. A landing zone becomes valuable when new workloads can inherit clear subscription boundaries, networking patterns, policies, and operational responsibility. Starting with a small team is an opportunity to make those choices comprehensible rather than copying an unnecessarily complex enterprise diagram.

On this page
  1. Separate the platform foundation from the workload
  2. Design management groups around governance differences
  3. Provide subscriptions through a governed process
  4. Make connectivity a standard but not an obstacle
  5. Apply policy as a versioned control system
  6. Define the shared responsibilities of platform and workload teams
  7. Test guardrails with one workload onboarding
  8. Evolve the foundation without fracturing it

Separate the platform foundation from the workload

An organization adopting Azure can create subscriptions quickly, but speed of creation is not the same as a sustainable operating model. Without agreed ownership, identity controls, connectivity, logging and policies, each new team invents its own foundation. The result may be several environments that work individually yet cannot be governed, supported or audited consistently. A small-team Azure landing zone is successful when its identity, networking, policy and cost controls remain maintainable, a practical concern throughout Microsoft AZ-104.

A landing zone is a deliberately prepared environment for Azure adoption. Microsoft’s Cloud Adoption Framework distinguishes the platform foundation from workload landing zones. The platform establishes shared governance and, when needed, centralized services; workload environments host the applications and services owned by delivery teams. Both are parts of one operating system, not competing architectures.

Consider a software product team requesting a new production environment. It should not have to negotiate basic log delivery, subscription ownership or allowed regions independently each time. Conversely, the central platform should not dictate every application component or prevent legitimate workload requirements without review. A landing zone succeeds when both the shared guardrails and the team’s freedom within them are explicit.

Design management groups around governance differences

Management groups let a platform team organize subscriptions for inherited policies and role assignments. The hierarchy should express governance differences that actually matter: platform services versus application workloads, regulated environments, or separate operational requirements. Copying the company’s department chart into the hierarchy can create needless moves and exceptions when teams reorganize.

A shallow, stable structure is generally easier to maintain than one with many levels and one-off branches. The design record should explain which standards apply at each level and why an exception belongs there. A subscription moved to a different management group can inherit a different policy set, so movement needs impact review rather than a casual drag-and-drop operation.

Use resource groups inside subscriptions for workload lifecycle boundaries, not as a substitute for subscription-level separation. Azure resource-group operations focus on changes and deletion scope within a subscription; a landing zone addresses broader governance and operating responsibility. Mixing those concepts is a frequent cause of ambiguous access and cost ownership.

Provide subscriptions through a governed process

A workload landing zone commonly begins with an approved subscription request. The request should identify environment, business owner, expected service, data classification, region constraints, billing accountability, connectivity needs and retirement conditions. Subscription vending automation can then apply known guardrails while leaving application configuration to the appropriate team.

The vending process should establish identity and management access, approved governance, baseline telemetry and any connectivity attachments required by the architecture. It should also record what automation created and how it can be updated safely. A human approval process that takes weeks to produce one manually configured subscription is likely to encourage shadow environments; a vending pipeline without an owner or rollback plan can create equally serious problems at scale.

A subscription is not just a cost bucket. Azure service limits, access boundaries, policy inheritance and certain operations are tied to its scope. Design subscription separation for meaningful operational concerns instead of either placing everything under one umbrella or creating a new subscription for every small component. Cost accountability should follow real service ownership, while shared platform spending needs an explicit allocation method.

Make connectivity a standard but not an obstacle

Application teams need predictable routes to shared DNS, on-premises systems, internet egress and other Azure workloads. A platform landing zone may provide hub-and-spoke networking, managed transit, central firewalls or resolver services depending on scale. The exact product choice matters less than ensuring that the traffic paths and security owners are understood.

A new subscription attached to central connectivity must receive more than an address range. Verify address-space coordination, routing, private name resolution, inspection paths, required service endpoints, and network change controls. When a private storage endpoint resolves differently in two spokes, the missing piece may be a shared DNS link, not a subscription misconfiguration. Network and identity layers still need distinct evidence during incidents.

Avoid a platform design that requires every team to request broad firewall exceptions for ordinary legitimate deployments. Publish approved patterns with clear request boundaries and testable defaults. Teams should know which services they may deploy directly, which require architecture review and how exceptions will expire. Security becomes easier to operate when the standard path is also the practical path.

Apply policy as a versioned control system

Azure Policy can audit or enforce requirements such as approved locations, diagnostic settings, security configurations and metadata standards, subject to feature support. Landing-zone policies should be deployed with version control and progressively tested; a sudden tenant-wide deny can interrupt necessary deployments and create pressure for uncontrolled exemptions.

Start with a documented objective rather than a collection of every built-in policy that sounds desirable. For example, restricting public exposure on certain storage accounts may be appropriate in production, but test how it affects deployment automation and approved exception workloads. Audit first where needed, fix noncompliance, then consider enforcement. A policy assignment is only as strong as the operational process that handles exceptions, remediation and evolving requirements.

Use Azure Policy initiatives to organize controls that share a real governance purpose; do not collect unrelated definitions simply for fewer assignment objects. Track exemptions separately from broad exclusions and make their owners and review dates clear. A landing zone is a living baseline, not a once-approved PDF of intended rules.

Define the shared responsibilities of platform and workload teams

A landing zone establishes a usable starting environment, but it cannot certify the reliability, security, cost or performance of every workload placed within it. The application owner should still conduct a Well-Architected review against the deployed service, including its failure modes, identity exposure, operating costs and recovery obligations. This separates the platform team’s responsibility for shared guardrails from an application team’s responsibility for actual service behavior, without leaving either group assuming that the other completed the work.

A landing zone can fail when every control has two owners or none. The platform team might own management-group policy, transit networking, core identity configuration and central log routing. An application team might own its workload identity assignments, data protection decisions, deployment pipelines and service health. The split should be documented per operational task and compatible with the organization’s governance model.

Take incident response as an example. When a production API loses access to storage, who checks private DNS? Who can change the storage firewall? Who decides whether an application secret or role assignment is compromised? An effective landing zone lets those teams identify and collaborate quickly. A central team that owns all operations can become a bottleneck; completely decentralized control can destroy consistent security evidence.

The same principle applies to lifecycle. Who retires idle test subscriptions? Who ensures that logs remain available for the required period? Which business owner accepts ongoing shared-service costs? Include decommissioning as a normal path from the start, rather than leaving old subscriptions unmanaged indefinitely.

Test guardrails with one workload onboarding

A useful pilot is to onboard a representative, nonproduction application using the intended subscription-vending path. Confirm the assigned management group, inherited policies, RBAC permissions, logging destination and network/DNS behavior. Deploy a small app and one supported dependent service, then attempt a deliberately disallowed configuration in a controlled test to verify that the expected policy guardrail works.

Review the experience from the workload team’s point of view. Could they perform approved work without unrecorded elevated access? Did the platform’s policy enforcement provide actionable error messages? Were documentation and support ownership clear? The cost of repeated manual exceptions is useful feedback that the design may not reflect actual development needs.

Run a second scenario in which the workload is retired. Determine what must be deleted, retained for compliance, or disconnected from shared services. This reveals why environment creation alone is insufficient as a landing-zone success metric. Managed infrastructure needs onboarding, operation and offboarding controls.

Evolve the foundation without fracturing it

As Azure use grows, new data services, regions, security controls and organizational boundaries will challenge the original design. Avoid rebuilding the management-group tree for every new team, and avoid freezing the baseline because changing it requires work. Version shared configurations, communicate changes, test representative workloads and retain a documented exception process that does not become permanent policy bypass.

A landing zone earns its value when the next application team can onboard without improvising its identity, connectivity, security, logging, or ownership model. Check whether a new workload inherits the intended baseline, whether exception requests have a real owner, and whether an incident responder can trace dependencies across subscription boundaries. The objective is repeatable safe operation, not a larger collection of diagrams and policies.

Continue learning

Related guides

Azure Management Groups at ScaleApply inherited governance across Azure management groups and subscriptions.Azure Subscription DesignSet subscription boundaries for ownership, billing, and policy.Subscriptions, Azure Policy, and Resource LocksSee how subscriptions, Azure Policy, and locks constrain deployments.

Related Posts

• Legacy MCSA Cloud Platform: Azure Architecture Skills and Current Paths

• Top Cloud Certifications Valuable in 2022: Is Microsoft a Leader?

• Microsoft AZ-104: Is Azure Administration the Right Career?

• Conquer the Cloud: Your Beginner’s Blueprint for Microsoft Azure Mastery

• Microsoft AZ-104: Six-Week Study Plan

• Azure Mastery Beyond the Basics: Governance, Green Cloud, and Strategic Growth

• Cloud Strategy Unlocked: How AZ-305 Shapes Real Azure Architects

• Cloud Confidence Starts Here: A Four-Part AZ-900 Certification Blueprint

• Microsoft AZ-104: Troubleshoot Azure VMs

• Microsoft AZ-104: Subscriptions, Policy, and Locks