Explore our new Cybersecurity Terms Glossary, explained by our experts.
Cybersecurity Terms Glossary
A managed identity gives an Azure resource its own Microsoft Entra ID identity, so applications can authenticate to Azure services without storing a secret anywhere.

What Are Azure Managed Identities?
A managed identity is an identity in Microsoft Entra ID that belongs to an Azure resource rather than to a person. A virtual machine, function app, container instance, or logic app can request an access token for an Azure service and be issued one on the strength of what it is, not on a client secret it has to keep. The platform creates the identity, rotates its credentials, and deletes it when the resource is deleted. Nothing lands in a configuration file, a pipeline variable, or a developer’s laptop.
Before managed identities, an application that needed to read a Key Vault secret or write to a storage account used an Entra ID application registration with a client secret or certificate. Those secrets had to be generated, distributed, stored, and rotated. In practice they were pasted into app settings, committed to repositories, and left to expire unnoticed. Cloud credential leakage is still one of the most common routes into a cloud tenant, and the fastest way to reduce it is to have fewer credentials in existence.
A system-assigned identity has a one-to-one relationship with a single resource. It is created when you enable it, shares the resource’s life cycle, and disappears when the resource is deleted. It suits a workload whose permissions should be unique to it.
A user-assigned identity is a standalone Azure resource that can be attached to many workloads. It survives independently, which makes it useful for scale sets, deployment slots, and blue-green rollouts where the compute is replaced regularly but the permissions should persist. The trade-off is sprawl: one identity shared by twenty services becomes a single key to twenty blast radiuses, and nobody remembers which workload still needs it.
A sensible default is system-assigned for anything with distinct permissions, user-assigned where a stable identity genuinely helps, and a naming convention that makes the owner obvious.
On a virtual machine, the workload calls the instance metadata service at the link-local address 169.254.169.254, supplies the required Metadata: true header, and receives a bearer token for the resource it asked for. App Service and Functions expose an equivalent local endpoint protected by an injected header value. The Azure Identity libraries wrap all of this behind DefaultAzureCredential, which is why most application code never sees the mechanics.
That convenience has a security consequence worth stating plainly: any code that can make an HTTP request from inside the resource can usually obtain the token. A server-side request forgery flaw in the hosted application, a compromised dependency, or a web shell all reach the same endpoint. This is the same class of exposure described in our entry on the IMDS attack, and it means a managed identity should be scoped as tightly as the application’s worst day, not its best.
RoleManagement.ReadWrite.Directory or AppRoleAssignment.ReadWrite.All can promote itself to Global Administrator.Sign-ins by managed identities appear in the Entra ID sign-in logs under the managed identity category, and resource access appears in Azure Activity Log and the data-plane logs of the target service. Route all of it into Microsoft Sentinel or your own platform and alert on a workload identity authenticating from an unexpected resource, requesting a token for a service it has never touched, or being granted a new role assignment outside change control. Good cloud logging and monitoring is what turns a silent identity into an auditable one.
Managed identities remove secrets; they do not remove the need to prove what each workload can actually reach. We trace workload identity paths across Azure subscriptions as part of our cloud security services.
The identity itself lives in Microsoft Entra ID, but the token endpoint is only available from inside an Azure resource that has the identity attached. For workloads running elsewhere, including Kubernetes clusters outside Azure, GitHub Actions, and other clouds, use workload identity federation instead. That lets an external token issuer be trusted directly, so you still avoid storing a long-lived client secret.
Yes. Anyone who can execute code or force an outbound request from the resource can usually request a token from the local metadata endpoint. The token inherits every role the identity holds. That is why scoping matters more than the absence of a stored secret: assume the identity will be borrowed at some point and grant it only what the application genuinely needs.
It is deleted with the resource, and its role assignments become orphaned entries that reference a principal ID no longer in the directory. Those stale assignments clutter access reviews and occasionally cause confusing audit findings. Clean them up as part of decommissioning, or use a scheduled check that removes role assignments pointing at principals that no longer exist.
No. A shared identity accumulates the union of every permission its consumers need, so each workload ends up with access it has no business holding. If one of them is compromised, the attacker inherits the whole set. Share a user-assigned identity only across instances of the same workload, and keep separate identities for separate applications and environments.
You May Be Interested In These
We use a few cookies that are needed to make this website work, and they are always on. We would also like to set analytics cookies, to see which pages are useful; functional cookies, so that our live chat works; and advertising cookies, to measure our campaigns. We will not set any of those until you say yes, and you can change your mind at any time from the Cookie Settings link in the footer.
Choose the cookies you are happy for us to use. Strictly necessary cookies are always on; everything else is off until you turn it on.
These cookies keep the website working: they hold your security checks, your form submissions and the choice you make here. The site cannot run without them, so they are always on.
These let our live chat open and remember your conversation, and let the newsletter sign-up work. Switch them off and the chat window will not appear.
These tell us which pages are read and how people find us, so that we can improve the site. The reports are about groups of visitors, never about you personally.
These let us measure our advertising and show you our adverts on other websites, such as LinkedIn, Google and Meta. Switch them off and you will still see adverts, but they will be less relevant.