What environments are, why they matter, and the foundational concepts every practitioner must understand.
What is a Power Platform Environment?
A Power Platform environment is a container that stores, manages, and shares your organization's business data, apps, flows, chatbots, and connections. Every Power Platform resource — whether a Canvas App, a Power Automate flow, or a Dataverse table — exists within exactly one environment.
Environments act as security, governance, and lifecycle boundaries. They control who can see what, where data lives, what policies apply, and how solutions move from development to production.
🔐 Security Boundary
Access to an environment is controlled by roles. Without an explicit role assignment, tenant users cannot access another environment's resources. Security groups can restrict who can even access the environment.
🔄 ALM Boundary
Solutions are exported from one environment and imported to another. A healthy ALM strategy uses separate Dev, Test/UAT, and Production environments. You cannot have ALM without environment discipline.
🔒 Governance Boundary
Data Loss Prevention (DLP) policies are applied at the tenant or environment level. Different environments can have different DLP policies, enabling you to permit premium connectors in controlled zones and restrict them elsewhere.
🌎 Data Residency
Each environment is tied to a geographic region. Data stored in Dataverse within that environment stays in the selected region. This is critical for compliance, data sovereignty, and regulatory requirements.
Common misconception: Many organizations treat the Default environment as a general-purpose workspace. The Default environment is accessible to all licensed users in the tenant and should never host sensitive business applications or production data.
Side-by-side view of all environment types across key attributes.
Default
Auto-created, tenant-wide
Production
Live workloads, real users
Sandbox
Dev, Test, UAT, training
Trial
Short-lived exploration
Developer
Personal development env
Dataverse for Teams
Lightweight, Teams-native — deprecated
| Attribute | Default | Production | Sandbox | Trial | Developer | Dataverse for Teams |
|---|---|---|---|---|---|---|
| Intended Purpose | Tenant fallback | Live workloads | Dev/Test/UAT | Exploration/PoC | Personal dev | Teams collaboration |
| Permanent? | Yes | Yes | Yes | No (30 days) | Yes* | Until Team deleted |
| Dataverse Supported? | Optional | Yes | Yes | Yes | Yes (included) | Limited subset |
| Good for Production? | No | Yes | No | No | No | No |
| Good for Development? | No | No | Yes | PoC only | Yes | Simple apps |
| Can be Reset? | No | No | Yes | No | No | No |
| Can be Copied? | No | Yes | Yes | No | No | No |
| Convert to Production? | No | N/A | Yes | Yes | No | Upgrade path |
| Managed Env Eligible? | Yes | Yes | Yes | No | No | No |
| Typical Owner | Tenant Admin | Business Unit / CoE | Dev Team / CoE | Individual | Individual Developer | Team Owner |
| Governance Level | Low (open) | High | Medium-High | Very Low | Low | Low-Medium |
| Full ALM Support? | No | Yes | Yes | Limited | Yes | No |
Detailed breakdown of every environment type — capabilities, limitations, admin controls, and real-world guidance.
Default Environment
The automatically created tenant-wide environment. Every licensed user has access.
Definition
Automatically created when the first user signs up for a Microsoft 365 or Dynamics 365 license in a tenant. There is exactly one Default environment per tenant. It cannot be deleted.
Who Creates It?
Automatically provisioned by Microsoft. No admin action required. Named "[TenantName] (default)".
Who Has Access?
Every licensed user in the tenant automatically has the Environment Maker role. This means any licensed user can create apps and flows here without explicit role assignment.
Dataverse
Dataverse can be added to the Default environment, but is not present by default. Once added, it cannot be removed without recreating the environment (which is impossible for Default).
Enterprise Risk: Because all licensed users are Environment Makers by default, the Default environment can quickly become a sprawl zone. This is the single most common Power Platform governance failure in enterprises.
What Makers Can Do
- Create Canvas Apps
- Create Power Automate flows
- Create Copilot Studio bots (verify licensing)
- Use standard connectors
- Share apps with colleagues
What Env Admins Can Do
- Manage environment settings
- Apply DLP policies at environment level
- Remove maker roles (with care)
- Enable Managed Environments
- Review analytics via admin center
What Tenant Admins Can Do
- Enable/disable Managed Environments
- Apply tenant-wide DLP policies
- Rename the environment
- Restrict who can be Environment Maker via tenant settings
- Block user access (via security group — Verify in current docs)
- Cannot be deleted or permanently removed
- Cannot be reset (unlike Sandbox)
- Cannot be copied to another environment
- All licensed users are automatically Environment Makers — cannot restrict without tenant-level policy changes
- Poor choice for ALM — no proper dev/test/prod lifecycle
- Managed Environments requires additional licensing (Verify in current Microsoft documentation)
- Dataverse capacity counts against tenant allocation
Governance Recommendation: In most enterprise deployments, the Default environment should be treated as a "citizen developer playground." Apply a restrictive DLP policy that blocks premium and sensitive connectors. Enable Managed Environments to gain usage insights and sharing controls.
Best For
- Exploring Power Platform capabilities
- Personal productivity flows (non-sensitive)
- Tenant-level integration scenarios
- Learning and training environments
Avoid When
- Building enterprise or departmental apps
- Storing sensitive or regulated data
- Requiring ALM, versioning, or deployment pipelines
- Apps need to be owned by a team, not an individual
Production Environment
For live business applications serving real users and real data.
Definition
A purpose-built environment for live workloads. Applications here serve real users, process real business data, and must meet uptime, security, and compliance requirements.
Who Creates It?
Power Platform Admin or Global Admin. Requires available environment capacity in the tenant (based on licensing). Creation can be restricted to admins via tenant policy.
Dataverse
Dataverse can be added at creation or later. With Dataverse, you gain full security model (business units, security roles, column-level security), backups, and enterprise ALM support.
Lifecycle
Permanent. Backed up automatically (daily and on-demand with Dataverse). Can be copied to a Sandbox. Cannot be reset. Convert a Sandbox to Production after promotion.
What Makers Can Do
- Build and publish Canvas Apps
- Create Model-driven Apps
- Build Power Automate flows
- Create Copilot Studio bots
- Use solutions for ALM
- Use premium and custom connectors
What Env Admins Can Do
- Manage security roles
- Create/restore backups
- Copy environment to Sandbox
- Manage environment settings
- Install Dynamics 365 apps (if licensed)
- Apply environment-level DLP overrides
What Tenant Admins Can Do
- Enable Managed Environments
- Apply tenant DLP policies
- Delete environment
- Transfer ownership
- View all environments in admin center
- Manage environment capacity
- Requires appropriate licenses for users (per-app, per-user, or Dynamics 365)
- Environment count limited by tenant license allocation (Verify in current Microsoft documentation)
- Cannot be reset without data loss
- Dynamics 365 apps require specific licensing
- Dataverse storage counts against tenant allocation
Best Practice: Never develop directly in Production. Always promote solutions from Sandbox through an ALM pipeline using Solutions and Deployment Pipelines or Azure DevOps / GitHub Actions.
Best For
- Enterprise applications with real users
- Business-critical workflows and automation
- Regulated data (with proper Dataverse security)
- Dynamics 365 workloads
- Solutions requiring SLA and uptime guarantees
Avoid When
- Building or testing — use Sandbox
- Proof of concept work — use Trial or Developer
- Short-term demos — use Trial
- Individual learning — use Developer or Default
Sandbox Environment
The essential environment for development, testing, UAT, and training.
Definition
A non-production environment designed for development, testing, user acceptance testing (UAT), and training. Can be reset to a clean state — ideal for test cycles and onboarding.
Key Differentiator
Sandbox is the only environment type that can be reset (wiped and rebuilt) and copied from a Production or another Sandbox. This makes it the cornerstone of ALM.
Can Be Converted
A Sandbox can be converted to a Production environment — enabling promotion of a fully tested environment without data re-migration. A key pattern in enterprise ALM.
Dataverse
Full Dataverse support. Backups can be taken and restored. Can receive copies from Production. Supports all solution types and Dynamics 365 apps (if licensed).
Maker & Dev Experience
- Full Canvas, Model-driven, Code app support
- Power Automate flows with all connector types
- Copilot Studio bots
- Premium and custom connectors
- Solution-based development (recommended)
- ALM via Deployment Pipelines
Admin Operations
- Reset environment (wipes all data and customizations)
- Copy from Production to Sandbox
- Convert Sandbox to Production
- Backup and restore
- Install Dynamics 365 apps (if licensed)
Governance Controls
- Assign security groups to restrict access
- Apply environment-specific DLP policies
- Enable Managed Environments
- Monitor with CoE Starter Kit
Recommended ALM Pattern: Developer Env (individual) → Sandbox Dev → Sandbox Test/UAT → Production. Solutions are exported from each stage and imported to the next via Deployment Pipelines or CI/CD automation.
Reset Use Case
When a sprint or release cycle completes, the Sandbox Dev environment can be reset and refreshed with a copy of Production — giving developers a clean, realistic dataset for the next cycle.
Copy Use Case
Before a major release, copy Production to a Sandbox UAT environment. Testers verify the new release against a near-identical copy of production data — without risking the live environment.
Best For
- Development (all types)
- Integration testing
- User Acceptance Testing (UAT)
- Training environments
- Performance testing with production data copies
Avoid When
- Serving real end users with live data
- Long-running stable integrations (use Production)
- Extremely cost-sensitive scenarios (Dataverse storage counts)
Trial Environment
Time-limited environments for exploration, PoC, and evaluation. Two types: Standard and Subscription-based.
Standard Trial (Self-Service)
- Created by any licensed user (if self-service allowed)
- Expires after 30 days
- Can be extended once (Verify in current Microsoft documentation)
- Dataverse included
- Cannot be converted to Production directly
Subscription-Based Trial
- Created from tenant-level trial subscription
- Typically 30 days tied to trial license
- Admin controlled — not self-service
- Can be converted to Production
- Full Dynamics 365 apps accessible
- More appropriate for formal PoC / evaluation
Architectural Warning: Trial environments should never be used for any workload that users will depend on beyond the evaluation period. When a trial expires, all data and apps in that environment are permanently deleted unless converted or extended.
Developer Environment
A personal Dataverse environment for individual developers. Not for production or shared workloads. Unmanaged developer environments (including those created by the free Power Apps Developer Plan) are deprecated: no new ones from 15 Nov 2026, existing ones suspended from 15 Dec 2026 and deleted from 15 Jan 2027. Going forward the Developer type exists only as a managed developer environment — one per user — and running apps/flows in it needs a premium license.
Key Facts
- Now created as a managed developer environment (PPAC, or automatically via environment routing) — the Developer Plan no longer provisions one from 15 Nov 2026
- Limit: one managed developer environment per user (it was up to three)
- Includes Dataverse with limited capacity (~1GB)
- Supports premium connectors for development
- Not intended for multi-user collaboration
What It Unlocks
- Full solution development and ALM tooling
- Dataverse tables, model-driven apps
- Power Automate with premium connectors
- Copilot Studio (verify licensing)
- Source control integration (Azure DevOps, GitHub)
Limitations
- Cannot be shared as a collaboration space
- Cannot be used for production workloads
- May be reclaimed if inactive (Verify in current Microsoft documentation)
- Limited Dataverse capacity
Managed Env Eligibility
Developer environments must now be managed. Unmanaged ones are being suspended (15 Dec 2026) and deleted (15 Jan 2027) — convert them by enabling Managed Environments (premium licensing required) or move their solutions to a sandbox. Personal developer environments created by environment routing are managed by default and unaffected.
Dataverse for Teams Environment (deprecated)
Dataverse for Teams environments are deprecated: no new ones can be created from 15 Nov 2026; environments not upgraded to Dataverse are suspended on 15 Sep 2027 and decommissioned on 30 Oct 2027. Build new Teams apps, flows and agents in a regular Dataverse environment and add them to Teams.
Creation
Automatically created when a Team owner creates their first Power App, flow, or Copilot Studio bot inside a Microsoft Teams channel using the Power Apps app in Teams.
Security Model
Mapped directly to the Microsoft Team. Team Owners get admin rights. Team Members are standard users. Team Guests have limited access. No traditional Environment Maker/Admin roles.
What It Supports
- Canvas Apps (Teams-embedded)
- Power Automate flows
- Dataverse for Teams tables
- Copilot Studio bots for Teams
- Standard connectors only
What It Does NOT Support
- Model-driven apps
- Premium connectors (without upgrade)
- Custom connectors
- Dynamics 365 apps
- Full Dataverse capabilities
- Full ALM via solutions
Upgrade Path (now mandatory to keep anything): Upgrade every Dataverse for Teams environment you still need to full Dataverse before 30 Oct 2027 (PPAC → Environments, filter Type = Microsoft Teams). The tenant needs free Dataverse capacity at least equal to the environment size, and users then need Power Apps / Power Automate / Copilot Studio licenses. This unlocks premium connectors, Model-driven apps, ALM, and the full Dataverse security model — but requires appropriate licensing.
Managed Environments
A premium governance layer — not a separate environment type. Enabled on top of existing environments.
Important Distinction: Managed Environments is NOT a standalone environment type. It is a premium governance capability that can be enabled on Default, Production, and Sandbox environments. It requires Power Platform Premium licensing (Verify in current Microsoft documentation).
Governance Features
- Weekly digest emails to admins
- Maker welcome content / onboarding
- Sharing limits (restrict sharing to security groups)
- Solution checker enforcement
- Data policies (additional controls)
- IP firewall (Verify current availability)
Admin Benefits
- Deep usage analytics per environment
- Understand who built what and when
- Enforce solution checker policies before deployment
- Visibility into connector usage
- Maker-facing guidance and guardrails
When to Enable
- Enterprise or departmental production environments
- Default environment governance
- CoE-governed Sandbox environments
- Environments with sensitive data
- When maker sprawl is a concern
What each persona can do. D365 = Requires Dynamics 365 license ✓ Full capability ~ Partial / conditional ✗ Not available
| Capability | App Maker | Env Maker | Env Admin | Sys Customizer | Sys Admin | PP Admin | Global Admin |
|---|---|---|---|---|---|---|---|
| Create Canvas Apps | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ |
| Create Power Automate Flows | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ |
| Create Dataverse Tables | ✗ | ✗ | ~ | ✓ | ✓ | ✓ | ✓ |
| Manage Dataverse Data (all records) | ✗ | ✗ | ~ | ~ | ✓ | ✓ | ✓ |
| Manage Environment Users | ✗ | ✗ | ✓ | ✗ | ✓ | ✓ | ✓ |
| Assign Dataverse Security Roles | ✗ | ✗ | ~ | ✗ | ✓ | ✓ | ✓ |
| Backup / Restore Environment | ✗ | ✗ | ✓ | ✗ | ✗ | ✓ | ✓ |
| Reset Environment | ✗ | ✗ | ✓ | ✗ | ✗ | ✓ | ✓ |
| Copy Environment | ✗ | ✗ | ✓ | ✗ | ✗ | ✓ | ✓ |
| Apply DLP Policies | ✗ | ✗ | ✗ | ✗ | ✗ | ✓ | ✓ |
| Enable Managed Environments | ✗ | ✗ | ✗ | ✗ | ✗ | ✓ | ✓ |
| Install D365 Apps | ✗ | ✗ | D365 | ✗ | ✗ | D365 | D365 |
| Use Premium Connectors | ~ License | ~ License | ~ License | ~ License | ~ License | ✓ | ✓ |
| Create & Manage Solutions | ~ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ |
| Run Deployment Pipelines (ALM) | ✗ | ~ | ✓ | ✗ | ✓ | ✓ | ✓ |
| Delete Environment | ✗ | ✗ | ✓ | ✗ | ✗ | ✓ | ✓ |
Key Distinction: Environment Admin can back up, copy, and reset environments, but cannot apply tenant-level DLP policies or enable Managed Environments — those require Power Platform Admin or Global Admin. System Administrator (Dataverse security role) has full control over Dataverse data and schema, but has no environment management capability outside Dataverse.
How Dataverse changes everything — roles, capabilities, security, and behavior.
Environments With vs. Without Dataverse
Without Dataverse
- Canvas Apps use SharePoint, SQL, Excel, or external data sources
- Power Automate flows run against external connectors
- No Dataverse security roles — only environment roles apply
- No model-driven apps possible
- No Dynamics 365 apps
- Limited governance and data management
With Dataverse
- Dataverse acts as the central data store within the environment
- Full security model: business units, security roles, teams, column-level security, field-level security
- Model-driven apps and Dynamics 365 apps become available
- Audit logging, encryption, data retention policies
- Backups managed at environment level
- Solutions with Dataverse table dependencies fully supported
Role Hierarchy: Critical Distinctions
Environment Roles
Assigned in Power Platform Admin Center. Control who can access and what they can build in the environment.
- Environment Admin — full env management
- Environment Maker — create resources (apps, flows)
Dataverse Security Roles
Assigned within Dataverse. Control access to data records, tables, and schema.
- System Administrator — full Dataverse control
- System Customizer — modify schema, no full data access
- Basic User — read own records
- Custom roles — any granularity
Common Mistake
Assigning Environment Admin does NOT give System Administrator in Dataverse — these are independent role systems. A user with Environment Admin but no Dataverse security role cannot read or modify Dataverse records. Both must be assigned explicitly.
Environment Maker grants the ability to create Power Platform resources within an environment. However, it does not grant any Dataverse security role. A user with only Environment Maker role can create a Canvas App that connects to a Dataverse table, but will receive an access denied error when the app tries to read data — unless they are also assigned a Dataverse security role with the appropriate table read permission.
For developers building model-driven apps or working directly with Dataverse tables, they need at minimum the System Customizer role (to create/modify tables) or a custom role with the appropriate privileges.
In Dataverse for Teams, security is inherited from the Microsoft Teams structure. Team Owners get elevated privileges, Team Members get standard access, and Team Guests get limited read access. Traditional Dataverse security roles and business units are not available.
In full Dataverse, the security model is significantly more granular — supporting business units, security roles, field-level security, record-level sharing, teams with role assignments, and column-level security. When upgrading a Dataverse for Teams environment to full Dataverse, a migration path is required to re-establish the security model.
Standard Trial (Self-Service)
| Created by: | Any licensed user (if self-service allowed) |
|---|---|
| Duration: | 30 days |
| Extendable? | Once (Verify current policy) |
| Convert to Prod? | No direct conversion |
| On expiry: | Environment + all data deleted |
| Best for: | Individual exploration, learning |
Subscription-Based Trial (Admin-Controlled)
| Created by: | Power Platform Admin or Global Admin |
|---|---|
| Duration: | 30 days (tied to trial subscription) |
| Extendable? | Via subscription extension (Verify) |
| Convert to Prod? | Yes — before expiry |
| On expiry: | Environment deleted unless converted |
| Best for: | Formal PoC, structured evaluation |
Architectural Anti-Pattern: Using a trial environment for a production workload is one of the most dangerous governance failures. When the trial expires, the environment and all data are deleted. Always migrate solution and data to a Production environment before the trial expires.
🆕 Why It Exists
Change (Nov 2026): Unmanaged developer environments (including those created by the free Power Apps Developer Plan) are deprecated: no new ones from 15 Nov 2026, existing ones suspended from 15 Dec 2026 and deleted from 15 Jan 2027. Going forward the Developer type exists only as a managed developer environment — one per user — and running apps/flows in it needs a premium license. Historically, the Developer Plan provided a free, personal development environment that allowed professional developers to build, test, and learn with Power Platform — including premium connectors and Dataverse — without consuming corporate environment capacity or requiring paid licenses for the development phase.
👥 Collaboration Implications
Developer environments are personal. While another user can be invited, the environment is designed for a single developer's workflow. For team-based development, use a dedicated Sandbox environment with proper access controls and shared solutions in source control.
📊 Capacity
Storage capacity is limited (~1GB Dataverse). Not suitable for large datasets. If the developer needs to work with realistic production-scale data, a Sandbox copy of Production is more appropriate.
✅ Perfect When
- Building and testing solutions before committing to Sandbox
- Learning Power Platform capabilities independently
- Developing connectors or integrations in isolation
- Contributing to open-source Power Platform community solutions
Lifecycle Tied to the Team
Dataverse for Teams environments are deprecated: no new ones can be created from 15 Nov 2026; environments not upgraded to Dataverse are suspended on 15 Sep 2027 and decommissioned on 30 Oct 2027. Historically, a Dataverse for Teams environment was created when the first Power App or flow was built inside a Microsoft Teams team. Its lifecycle is directly tied to the Team — if the Team is deleted, the associated environment and all its data are also deleted after a retention period.
Ideal Use Cases
- Simple team productivity apps
- Departmental data collection (non-critical)
- Small-scale automation for Teams workflows
- Rapid prototyping by business users
- Apps used exclusively within one Microsoft Team
Warning Signs — Time to Upgrade
- App needs to be used outside the Teams context
- Data volume approaching capacity limits
- Need for model-driven apps or Dynamics 365
- Need for premium connectors
- Need for advanced security roles
- Need for full ALM pipeline
Sharing Controls
Restrict Canvas App sharing to security groups only. Prevent makers from sharing apps broadly across the tenant. Enforce maker onboarding — makers see custom welcome content when they enter a Managed Environment for the first time.
Solution Checker Enforcement
Require solutions to pass the Solution Checker (which identifies security, reliability, and performance issues) before they can be deployed. This catches issues early and improves solution quality across the organization.
Weekly Admin Digest
Admins receive a weekly email summarizing new apps, flows, makers, and connections created in the environment. This provides passive visibility without requiring active monitoring — a critical tool for governance at scale.
Licensing Requirement
Managed Environments require Power Platform Premium per-user or per-flow licensing, or Managed Environments standalone licensing. Verify current licensing requirements in official Microsoft documentation.
Proven patterns for common organizational scenarios.
👤 Solo Maker Learning Path
Use Default or Developer environment for learning. When building something real, request a dedicated Sandbox.
Low GovernanceDeveloper Plan📂 Small Departmental App
One Sandbox for development, one Production. Restrict access via security groups. Use solutions for deployment.
Sandbox → ProductionSolutions🏢 Enterprise ALM Strategy
Dev → Test → UAT → Production pipeline. CI/CD via Azure DevOps or GitHub Actions. Managed Environments on Production.
Full ALMManaged Envs👥 Teams-First Lightweight Solution
Don't start in Dataverse for Teams any more — creation ends 15 Nov 2026. Use a Developer or Production Dataverse environment and surface the app in Teams; upgrade any existing Teams environments before 30 Oct 2027.
Teams EnvUpgrade Path Planned📊 Proof of Concept
Use subscription-based Trial or Developer environment. Set a deadline. Convert or export before expiry.
Trial / Dev EnvTime-boxed🛠 Regulated Production Workload
Production with Managed Environments enabled. Security groups restrict access. DLP policy. Audit logging. BU-level security. Full ALM pipeline.
ProductionManaged EnvsDLPAnswer the question below to get an environment recommendation.
Recommendation
✅ Best Practices
Use Solutions for All Deployments
Never deploy individual components. Always use Solutions — they capture dependencies and enable proper ALM.
Separate Dev, Test, and Production
At minimum, maintain a Sandbox for development/testing and a Production environment for live users. Promote solutions via Solutions.
Apply Security Groups to Environments
Restrict who can access each environment by associating a Microsoft Entra security group. Prevents unauthorized access and sprawl.
Use Managed Environments on Production
Enable Managed Environments on production and default environments to gain visibility, enforce policies, and get weekly admin digests.
Plan Environment Naming Conventions
Use consistent naming: [Org]-[Workload]-[Stage] (e.g., Contoso-HRApps-Prod). This enables easier governance at scale.
Apply DLP Policies by Environment Tier
Default and non-governed environments should have restrictive DLP. Production environments for approved workloads can have specific connector permissions scoped to their needs.
⚠ Anti-Patterns & Common Mistakes
Using Default for Enterprise Apps
The Default environment has no meaningful access control by default. Every licensed user is an Environment Maker. Never host real business apps here.
Developing Directly in Production
Changes made directly in Production are untested, unversioned, and risky. Even small changes can break live apps for all users. Always use a Sandbox.
Relying on Trial for Real Workloads
When a trial expires, the environment and all data are permanently deleted. This has caused real data loss incidents. Never use Trial for anything users depend on.
Ignoring Dataverse Security Role Differences
Assigning "Environment Admin" without a Dataverse security role leaves users unable to access Dataverse data. These are two separate security systems.
Leaving Dataverse for Teams environments un-upgraded
Dataverse for Teams is deprecated — environments that aren't upgraded are suspended on 15 Sep 2027 and decommissioned on 30 Oct 2027, taking their apps, flows, agents and data with them. Inventory, classify and upgrade the ones you need now.
Not Planning ALM from Day One
Retrofitting ALM into a mature Power Platform deployment is extremely difficult. Design your environment strategy before the first app goes live.
Architect-level guidance for enterprise Power Platform environment strategy.
🌎 Regional Considerations
Choose environment regions based on data residency requirements. For EU/UK regulated data, select EU or UK regions. For latency-sensitive apps, choose a region close to users. Region cannot be changed after environment creation.
👥 Security Group Strategy
Associate one Microsoft Entra security group per environment. Use nested groups for role hierarchy. Example: PP-ENV-HRApps-Prod (parent) containing sub-groups for Admins, Makers, and Users.
🔒 DLP Strategy
Apply a tenant-wide baseline DLP policy blocking all premium connectors by default. Then create environment-specific override policies for approved production environments. The most restrictive applicable policy wins.
📄 Solution Strategy
Use one solution per business application or workload. Use publisher prefixes consistently (e.g., contoso_). Use layered solutions: base components + customization layer. Store solution source in version control.
⚖ Centralize vs. Decentralize
Centralized (CoE model): Admin team owns all environments. Best for regulated industries. Decentralized: Business units own their environments with central oversight. Best for agile organizations. Hybrid is most common at scale.
🏢 Business Unit Isolation
Isolate high-sensitivity business units via separate environments with their own DLP policies and security groups. Use separate environments for HR, Finance, and Legal workloads rather than sharing one Production environment.
Format: [OrgPrefix]-[Workload/Department]-[Stage]
- Contoso-HRApps-Dev
- Contoso-HRApps-Test
- Contoso-HRApps-Prod
- Contoso-Finance-Sandbox
- Contoso-Default (the tenant default)
- Contoso-CoE-Governance (CoE Starter Kit environment)
Consistent naming enables filtering, reporting, and policy management in the Power Platform Admin Center and CoE Starter Kit at scale.
Update (2026): the CoE Starter Kit is no longer actively maintained. Its core capabilities are now native in the Power Platform admin center — Inventory (all apps, flows and agents across the tenant), Usage, Monitor and Actions — so start there. The Power Platform Center of Excellence (CoE) Starter Kit is a set of apps, flows, and dashboards that provide visibility and governance across your entire tenant. It shows which apps and flows exist in which environments, who created them, when they were last modified, and what connectors they use.
Historically, without the kit admins were essentially blind to what was happening across environments. Today PPAC Inventory covers that natively. Only deploy the kit — into a dedicated CoE Governance environment (typically a Sandbox) — for gaps PPAC doesn't fill yet (maker surveys, bulk permission updates, abandoned-resource cleanup, ROI tracking), and treat it as unsupported code you own.
Mandatory settings, security controls, and ALM configuration every production environment must have — covering both Power Platform Pipelines and Azure DevOps approaches.
🛡️ Part 1 — Mandatory Production Governance & Security Settings
These are not optional. Each of these settings either prevents data loss, enforces access control, or enables rollback. Skipping any of them creates a governance or security gap that auditors, security teams, and the CoE will flag.
1. Enable Managed Environments
Where: Power Platform Admin Centre → Environments → [Prod Env] → Enable Managed Environment
Managed Environments activates the governance layer on top of your production environment. It is the single most impactful switch you can enable on a production environment.
- Sharing limits for Canvas Apps (e.g. max 20 users, share only to security groups)
- Solution checker enforcement on every solution import
- Weekly digest email to admins with usage insights
- Maker welcome message — makers see your governance guidance on first use
- Usage insights — see which apps have no users (orphan detection)
- IP firewall and IP cookie binding (additional security controls)
- Requires Power Platform Premium (per-user) or Managed Environment add-on per environment
- Users of apps in a Managed Environment need Power Apps Premium or per-app licences
- Always validate current licensing at aka.ms/PowerPlatformLicensing
2. Restrict Environment Access with a Security Group
Where: PPAC → Environments → [Prod Env] → Edit → Restrict access to security group
Without a security group assignment, every licensed user in the tenant can access the production environment. This is a major governance gap.
| Environment | Security Group | Members |
|---|---|---|
| Production | SG-PP-Prod-Users | Business users who need app access |
| UAT | SG-PP-UAT-Testers | Testers, business owners, CoE lead |
| Dev | SG-PP-Dev-Team | Developers and makers only |
| CoE Governance | SG-PP-CoE-Admins | Platform admins and CoE leads only |
3. Apply a Hardened DLP Policy
Where: PPAC → Policies → Data Policies → Create or assign environment policy
Production environments need the most restrictive DLP policy. Use a layered approach: tenant baseline policy + environment-specific hardening policy.
| Group | What Goes Here (Production) |
|---|---|
| Business | SharePoint, Dataverse, Teams, Outlook, OneDrive for Business, Microsoft Entra ID, Approvals, SQL Server (if approved) |
| Non-Business | Low-risk, non-data-sensitive connectors only — typically very few in a hardened prod policy |
| Blocked | All consumer connectors: Gmail, Dropbox, Twitter/X, Facebook, Instagram, Google Sheets, personal OneDrive, RSS, and all unreviewed third-party connectors |
Most restrictive wins: When both a tenant-level and environment-level policy apply, the combination that is most restrictive is enforced. Design your tenant policy as a safety floor and your prod policy as a hardened ceiling.
4. Enable Connector Endpoint Filtering
Where: PPAC → Policies → Data Policies → [Policy] → Connector configurations
Endpoint filtering lets you restrict a connector to specific URLs or endpoints. For example: allow SharePoint connector only to https://yourorg.sharepoint.com — blocking connections to external SharePoint tenants.
| Connector | Endpoint Pattern | Rule Type |
|---|---|---|
| SharePoint | https://yourorg.sharepoint.com/* | Allow (deny all others) |
| SQL Server | yourdbserver.database.windows.net | Allow (deny all others) |
| HTTP with Microsoft Entra ID | https://api.yourorg.com/* | Allow (deny all others) |
5. Configure Tenant Isolation
Where: PPAC → Policies → Tenant Isolation
Tenant isolation prevents connectors in your tenant from authenticating to other organisations' tenants. This stops data being silently routed to a personal or external tenant via an authenticated connector. Enable for all production environments in regulated industries.
- Inbound restriction: Blocks makers from external tenants connecting to your tenant's resources
- Outbound restriction: Blocks your makers from connecting to another organisation's tenant resources
- You can add explicit allow-list exceptions for trusted partner tenants
6. IP Firewall & IP Cookie Binding (Managed Environments)
Where: PPAC → Environments → [Prod Env] → Settings → IP Firewall / IP Cookie Binding
Restricts access to the environment to requests originating from specified IP ranges. Enforces network-level access control — users outside approved networks (e.g. corporate VPN or office IP ranges) cannot access production, even if authenticated.
Binds session cookies to the IP address used at login. Prevents session cookie theft and replay attacks — a stolen cookie cannot be used from a different IP address. Recommended for all production Dataverse environments handling sensitive data.
7. Enable Audit Logging & Monitoring
Where: PPAC → Environments → [Prod Env] → Settings → Auditing
| Setting | Where to Enable | Why |
|---|---|---|
| Dataverse audit logging | PPAC → Environment Settings → Auditing → Start Auditing | Records every create/update/delete on Dataverse records — required for compliance and investigation |
| User access auditing | PPAC → Environment Settings → Auditing → Log access | Records who accessed what data and when |
| Microsoft Purview / M365 Unified Audit Log | M365 Admin Centre → Compliance → Audit | Cross-platform audit trail including Power Apps activity, flow runs, and data access |
| PPAC Inventory / Usage monitoring (CoE Starter Kit only for gaps) | CoE environment — App & Flow Analytics | Visibility into which apps are active, unused, or high-risk |
| PPAC Analytics | PPAC → Analytics → Capacity / Usage | Environment capacity, storage, and usage trends |
8. Enforce Solution Checker on Import (Managed Environments)
Where: PPAC → Environments → [Prod Env] → Managed Environment settings → Solution checker enforcement
When Solution Checker enforcement is enabled, every solution import to this environment is automatically scanned. You can configure the policy to warn (allow import but show issues) or block (reject imports with critical issues). Set production to block on critical issues.
This is a quality gate that prevents low-quality or security-risky solutions from reaching production. Combined with pipeline-based pre-deployment checks, it creates defence-in-depth for solution quality.
🔁 Part 2 — ALM with Power Platform Pipelines
Power Platform Pipelines is the native low-code deployment feature. It promotes solutions from Dev → Test → UAT → Prod through a configured pipeline without needing Azure DevOps. Best for teams with limited DevOps experience or for faster initial ALM setup.
Pipeline Setup — Step by Step
PPAC → Environments → New → type: Production or Sandbox. This environment hosts the pipeline configuration. It requires Dataverse. Name it clearly: Contoso-PipelineHost.
From the pipeline host environment, go to Power Platform Pipelines → Install the Deployment Pipeline Configuration app from AppSource into the host environment.
In the Deployment Pipeline Configuration app: add each environment (Dev, Test, UAT, Prod) as a registered environment. The host environment connects to them via the System Administrator role.
Create a new Pipeline → Add stages in order: Dev (source) → Test → UAT → Production. Each stage maps to a registered environment. Enable Pre-deployment approvals on UAT and Production stages.
For each target environment, configure the Deployment Configuration record — mapping connection references to actual connections and setting environment variable values specific to that target. This replaces the manual deployment settings JSON used with PAC CLI.
In the Dev environment, open the solution → Deploy → Select pipeline → Select target stage. The pipeline exports managed, validates, and imports. Approvers are notified. The solution promotes automatically to the next stage on approval.
| Pipeline Capability | Available | Notes |
|---|---|---|
| Pre-deployment approval | ✔ | Per-stage — approver receives Teams/email notification |
| Managed solution export | ✔ | Automatic — pipeline always deploys managed |
| Deployment history | ✔ | Full history in the pipeline host environment |
| Connection reference mapping | ✔ | Via Deployment Configuration in the host environment |
| Environment variable mapping | ✔ | Via Deployment Configuration in the host environment |
| Source control integration | ✘ | Not built-in — use Azure DevOps for Git-based ALM |
| Automated testing | ✘ | No built-in test runner — use Azure DevOps for this |
| Custom deployment steps | ~ | Limited — extensibility via pipeline-related flows |
⚙️ Part 3 — ALM with Azure DevOps
Azure DevOps provides enterprise-grade ALM: Git source control, branch strategies, PR validation, automated testing, solution checker enforcement, and full audit trail. Use it when your organisation has Azure DevOps licences and needs Git-based ALM, automated quality gates, or multi-solution dependency management.
Prerequisites to Configure
- Azure DevOps Organisation created
- Project created per domain/workstream
- Git repository initialised
- Power Platform Build Tools extension installed from Visual Studio Marketplace
- Service connections created (one per environment)
- Service principal (App Registration) created in Microsoft Entra ID
- Service principal added as System Administrator or Deployment role in each target environment
- Client secret generated and stored in Azure DevOps Library Variable Group (marked secret)
- PAC CLI available on pipeline agent (auto-installed via Build Tools task)
Service Connection Setup (per environment)
In Azure DevOps: Project Settings → Service Connections → New → Power Platform (added by Build Tools extension)
| Field | Value |
|---|---|
| Connection name | PP-Dev-Connection, PP-Test-Connection, PP-Prod-Connection |
| Server URL | https://yourorg.crm11.dynamics.com (environment URL) |
| Tenant ID | Your Microsoft Entra ID tenant ID (from Azure portal) |
| Application (client) ID | App registration client ID (service principal) |
| Client secret | Secret from App Registration — store in Variable Group, reference as $(ClientSecret) |
Recommended Pipeline Architecture (YAML)
Trigger: PR to main
Export unmanaged → unpack → commit to Git → run Solution Checker
Trigger: merge to main
Pack managed → deploy to Test → run automated checks
Trigger: manual / release tag
Deploy to UAT (approval) → deploy to Prod (approval)
# azure-pipelines-deploy.yml — CD to Production
trigger: none # manual trigger only for prod
pool:
vmImage: 'windows-latest'
variables:
- group: PP-Prod-VarGroup # holds ClientSecret, TenantId, ClientId
stages:
- stage: DeployToUAT
jobs:
- deployment: DeployUAT
environment: 'PowerPlatform-UAT' # Azure DevOps Environment with approval
strategy:
runOnce:
deploy:
steps:
- task: PowerPlatformToolInstaller@2
displayName: 'Install PAC CLI'
- task: PowerPlatformImportSolution@2
displayName: 'Import to UAT'
inputs:
authenticationType: 'PowerPlatformSPN'
PowerPlatformSPN: 'PP-UAT-Connection'
SolutionInputFile: '$(Pipeline.Workspace)/drop/ACME_HR_managed.zip'
UseDeploymentSettingsFile: true
DeploymentSettingsFile: '$(Build.SourcesDirectory)/settings/deploymentsettings.UAT.json'
- stage: DeployToProd
dependsOn: DeployToUAT
jobs:
- deployment: DeployProd
environment: 'PowerPlatform-Production' # Azure DevOps Environment with approval gate
strategy:
runOnce:
deploy:
steps:
- task: PowerPlatformToolInstaller@2
displayName: 'Install PAC CLI'
- task: PowerPlatformImportSolution@2
displayName: 'Import to Production'
inputs:
authenticationType: 'PowerPlatformSPN'
PowerPlatformSPN: 'PP-Prod-Connection'
SolutionInputFile: '$(Pipeline.Workspace)/drop/ACME_HR_managed.zip'
UseDeploymentSettingsFile: true
DeploymentSettingsFile: '$(Build.SourcesDirectory)/settings/deploymentsettings.Prod.json'
Deployment Settings File — deploymentsettings.Prod.json
Store one file per environment in source control. This file maps environment variables and connection references at import time — the solution itself stays environment-agnostic.
{
"EnvironmentVariables": [
{
"SchemaName": "acme_SharePointSiteURL",
"Value": "https://yourorg.sharepoint.com/sites/ProdHR"
},
{
"SchemaName": "acme_ApprovalTimeoutHours",
"Value": "48"
}
],
"ConnectionReferences": [
{
"LogicalName": "acme_sharedsharepoint_xxxxxxxx",
"ConnectionId": "/providers/Microsoft.PowerApps/apis/shared_sharepointonline/connections/prod-sp-svc-conn",
"ConnectorId": "/providers/Microsoft.PowerApps/apis/shared_sharepointonline"
},
{
"LogicalName": "acme_sharedapprovals_xxxxxxxx",
"ConnectionId": "/providers/Microsoft.PowerApps/apis/shared_approvals/connections/prod-approval-conn",
"ConnectorId": "/providers/Microsoft.PowerApps/apis/shared_approvals"
}
]
}
Find the LogicalName of a connection reference in the solution XML (connectionreferences.xml after unpacking). Find the ConnectionId in PPAC → Connections → click the connection → the URL contains the ID.
🔍 ALM Tool Decision Matrix
| Requirement | PP Pipelines | Azure DevOps | Both |
|---|---|---|---|
| Quick initial ALM setup (days) | ✔ Best | ~ | — |
| No DevOps skills in the team | ✔ Best | ✘ | — |
| Git-based source control | ✘ | ✔ Best | — |
| PR validation and code review | ✘ | ✔ Best | — |
| Automated solution checker in CI | ✘ | ✔ Best | — |
| Automated testing in pipeline | ✘ | ✔ Best | — |
| Multi-solution dependency chain | ~ | ✔ Best | — |
| Maker self-deployment with approvals | ✔ Best | ~ | — |
| Enterprise full ALM maturity (L4/L5) | ~ | — | ✔ Best |
✅ Production ALM Readiness Checklist
- All production components are inside named, versioned solutions with a custom publisher prefix
- Solutions are exported as managed before deploying to Test/UAT/Prod — never unmanaged
- Environment variables used for all environment-specific URLs, IDs, and configuration values
- Connection references in the solution — all flows use connection references, not direct connections
- Deployment settings file created and stored in source control for each target environment
- Service account or service principal used for all production flow connections — no personal accounts
- Solution Checker passes with no critical or high issues before import to Test or above
- Pre-deployment approval gate configured on UAT and Production pipeline stages
- Post-deployment validation steps documented and executed after every production release
- Rollback procedure documented — prior managed solution .zip retained for each release
- Release notes written with every deployment: version, changes, deployer, date
- Production flows turned on and verified after managed solution import
- No direct edits made in UAT or Production — all changes originate in Dev and flow through the pipeline
I want to explore Power Platform
Use your managed Developer environment (one per user — the free Developer Plan no longer creates environments from 15 Nov 2026) or the Default environment for simple experiments.
I need a time-limited proof of concept
Use a Subscription-based Trial environment. Plan a conversion path if the PoC succeeds.
I need to build and test a solution
Use a Sandbox environment. Separate Dev and Test environments for mature ALM.
I need to deploy for real users
Use a Production environment. Enable Managed Environments. Restrict via security group.
My app only needs to work in Teams
Build it in a regular Dataverse environment and embed it in Teams — Dataverse for Teams is deprecated and can't be created from 15 Nov 2026.
I'm a developer building solutions solo
Use your managed Developer environment (one per user, governed by Managed Environments; you need a premium license to run what you build there).
I need full ALM and enterprise governance
Dev Sandbox → Test Sandbox → Production, with Managed Environments, DLP policies, and CI/CD via Deployment Pipelines.
I need a training environment
Use a Sandbox that can be reset after each training cohort — delivering a clean state every time.
✅ Final Governance Checklist
- Every environment has a named owner and a defined purpose
- Security groups are assigned to restrict environment access
- DLP policies are applied — at minimum, the tenant-wide baseline
- Production environments have Managed Environments enabled
- All solutions use a consistent publisher prefix and naming convention
- No production apps exist in the Default environment
- ALM pipelines exist for all production-bound workloads
- Trial environments have a defined expiry and conversion plan
- Developer environments are not used for shared or production workloads
- Every Dataverse for Teams environment is either upgraded to Dataverse or signed off for decommission before 30 Oct 2027
- Environment Admin and System Administrator roles are assigned separately
- Region is correctly set for data residency compliance
- Environment sprawl is monitored via PPAC Inventory / Usage (or the legacy CoE Starter Kit)
Disclaimer: Licensing, capacity limits, feature availability, and role behavior in Power Platform can change with Microsoft product updates. Always verify the details in this guide against the current official Microsoft documentation before making architectural decisions. Specific numbers (storage limits, trial durations, role names) are subject to change.
- 🔗 Microsoft Learn: Types of Power Platform environments — learn.microsoft.com/power-platform/admin/environments-overview
- 🔗 Microsoft Learn: Create and manage environments — learn.microsoft.com/power-platform/admin/create-environment
- 🔗 Microsoft Learn: Dataverse security model — learn.microsoft.com/power-platform/admin/wp-security
- 🔗 Microsoft Learn: Managed Environments overview — learn.microsoft.com/power-platform/admin/managed-environment-overview
- 🔗 Microsoft Learn: Deprecation of unmanaged developer environments — learn.microsoft.com/power-platform/important-changes-coming
- 🔗 Microsoft Learn: Dataverse for Teams — learn.microsoft.com/power-apps/teams/overview-data-platform
- 🔗 Microsoft Learn: DLP policies — learn.microsoft.com/power-platform/admin/wp-data-loss-prevention
- 🔗 Microsoft Learn: ALM with Power Platform — learn.microsoft.com/power-platform/alm/overview-alm
- 🔗 Microsoft Learn: Power Platform CoE Starter Kit — learn.microsoft.com/power-platform/guidance/coe/starter-kit