Explore our new Cybersecurity Terms Glossary, explained by our experts.
Cybersecurity Terms Glossary
Google Cloud organisation policies apply constraints across the resource hierarchy, restricting what can be configured in every project beneath them regardless of local IAM rights.

What Are GCP Organisation Policies?
An organisation policy in Google Cloud is a constraint applied to a node in the resource hierarchy—the organisation, a folder, or a project—that restricts what can be configured anywhere beneath it. Unlike an IAM policy, which decides who may act, an organisation policy decides what may exist. A project owner with full IAM rights still cannot create a resource the policy forbids, which is what makes it a guardrail rather than a permission.
Google Cloud resources sit in a tree: organisation, then folders, then projects, then the resources themselves. Policies set at a node are inherited downwards, which means the structure of your folders determines how cleanly you can apply controls. Folders that reflect trust and environment—production, non-production, regulated, sandbox—let you write one policy per concern. Folders that mirror a reorganising business unit chart do not, and you will be rewriting exceptions for years.
Inheritance can be overridden at a lower node where the constraint permits it, which is how you grant a documented exception to a single project. Boolean constraints can be enforced or not enforced; list constraints allow or deny specific values, with an option to merge with or replace the inherited list.
iam.disableServiceAccountKeyCreation—blocks the long-lived JSON key files that are the most common Google Cloud credential leak. Pair it with workload identity federation so teams have an alternative.storage.publicAccessPrevention—stops buckets being made publicly readable, enforced regardless of bucket-level settings.sql.restrictPublicIp—keeps Cloud SQL instances off the public internet.compute.vmExternalIpAccess—restricts which virtual machines may have an external address, usually to an explicit allowlist.gcp.resourceLocations—confines resource creation to approved regions, which supports the data residency position you will have to defend.iam.allowedPolicyMemberDomains—prevents IAM grants to identities outside your own Cloud Identity domains, closing an easy route to external sharing.compute.requireOsLogin—forces SSH access through IAM-managed OS Login rather than project-wide metadata keys.compute.requireShieldedVm—requires secure boot and integrity monitoring on new instances.These eight close a large share of the weaknesses an assessment typically finds in a Google Cloud estate, and none of them costs anything.
Where the built-in list does not cover your requirement, custom organisation policy constraints let you write a condition in Common Expression Language against a resource’s fields and enforce it at creation or update. Typical examples include requiring a specific label, forbidding a machine type, or demanding customer-managed encryption keys on a particular service. Support varies by resource type, so check the documentation for the specific service before designing a control around it.
Organisation policies support a dry-run mode that evaluates and logs violations without blocking anything. Use it every time. Set the policy in dry run at the target node, review the audit log entries for would-be denials across a full release cycle, fix or exempt what appears, then enforce. Skipping this step is how a constraint that looks obviously safe discovers the one pipeline that creates a service account key every night.
Beyond that, keep policies in version control and deploy them with Terraform or the Cloud Deployment Manager, so a change is reviewed and reversible. Record every exception with an owner and an expiry, and review the exception list quarterly, because an override added for a migration tends to outlive the migration by several years.
Organisation policies constrain resource configuration. They do not evaluate who may call an API, which remains the job of IAM and the role bindings beneath it, and they do not see runtime behaviour. A project where every constraint is satisfied can still hold a service account with the Owner role, an overly trusting workload identity pool, or an application with a straightforward vulnerability.
Combine them with least-privilege IAM, Security Command Centre for posture findings, VPC Service Controls where you need a perimeter around sensitive services, and audit logging routed to a dedicated project as described in cloud logging and monitoring. The equivalent controls in the other providers are covered in AWS service control policies and Azure Policy.
At Privia Security, we review Google Cloud hierarchies and test the guardrails on them through our cloud security services.
IAM decides who may perform an action; an organisation policy decides whether a configuration is permitted at all. A project owner with every IAM role still cannot create a resource that a constraint forbids. That separation is what makes constraints useful as central guardrails: they survive delegation, because the people you delegated to cannot remove them from below.
Dry run evaluates a constraint and records violations in the audit log without blocking anything. Use it for every new policy, at the node where you intend to enforce, for at least one full release cycle. It surfaces the scheduled jobs and infrastructure pipelines that would break, which is almost always more than the team expects, before anything is actually denied.
Disabling service account key creation usually gives the largest single improvement, because those long-lived JSON key files are the most common route to a compromised Google Cloud project. Provide workload identity federation as the alternative before you enforce it, or you will simply generate a queue of exception requests that undoes the benefit.
Constraints are evaluated when a resource is created or updated, so existing non-compliant resources keep running. Use dry-run logging and Security Command Centre findings to build a list of what is already out of line, remediate through normal change control, and only then enforce. Some constraints, such as public access prevention, do take effect on existing resources, so check each one.
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.