Network, logging and monitoring

Are development, testing, and production conducted in separate environments?

The real question underneath is whether production data has been copied into environments with weaker controls.

Partly verifiable from your tenant

What the carrier is actually asking

The carrier is asking whether development, test, and production run in separate environments with separate access. It is asking about isolation and, implicitly, about whether real data is used outside production.

Why it is underwritten

Non-production environments routinely have weaker access control, no monitoring, and broader developer access, and they routinely contain a copy of production data. That combination has produced a long list of breaches where the compromised system was a test environment nobody was watching.

Where the answer lives in Microsoft 365, Entra ID, and Azure

In the cloud, separation is measurable through subscription and identity boundaries.

PlatformWhere the setting livesWhat has to be true
AzureSeparate subscriptions or management groups per environmentEnvironments in distinct scopes with distinct role assignments
AzureRole assignments per environmentProduction access is a separate grant, so development access does not confer it
AzureNetwork connectivity between environmentsNo routed path from development into production data services
Data handlingWhether production data is copied into lower environmentsMasking or synthetic data, since the alternative extends the regulated footprint. Attested
AzurePublic exposure of non-production resourcesTest systems are frequently the ones left publicly reachable, so check them specifically
Test data is production data

If a copy of the customer database sits in a test environment, that environment is inside the scope of every breach and every privacy obligation, with none of the controls. Masking is the fix, and it is deferred more often than any other data control.

What a defensible yes requires

  • Environments are separate at the subscription or account level with distinct access.
  • Production access is granted separately and held by fewer people.
  • Non-production data is masked or synthetic.
  • No network path connects development directly to production data.
  • Non-production resources are checked for public exposure, since they are the most likely to have it.

How this answer goes wrong

Environments are separated by naming convention inside one subscription, with one role assignment covering all of them. Or they are properly separated and the test environment holds a full copy of production data with developer-wide access.

Frequently asked

Is separation by resource group enough?

It is weaker than subscription separation because inherited permissions cross it easily. It can work with careful role design, and subscription separation is the cleaner boundary.

Can we use production data in testing?

Only masked or with the same controls as production. Most organisations underestimate how much regulated data sits in their test environments.

What about developer access to production?

It should be exceptional, time-bound, and logged. Standing developer access to production is a common finding and a real exposure.

Does this apply to software-as-a-service?

Yes, where sandbox instances hold real data. The same principle applies to platforms you do not host.

Related questions

Stop answering this from memory

Connect Microsoft 365, Entra ID, and Azure read-only. Insurance Posture reads the live configuration behind each application answer and shows you which ones you can prove before you sign.

Assess your posture