Resource Quotas

Every organization has a quota profile that caps how many of each kind of resource it can hold. The platform checks it whenever you create, import or scale up something, and refuses the change if it would go over a limit.

How quotas work

• Per organization: all projects in an organization draw from the same limits.
• Counts, not capacity: quotas limit the number of objects, not CPU, memory or storage. A large virtual machine uses one unit, just like a small one.
• Set by your platform operator: new organizations start on the platform's default profile (for example, Free). You can't change your organization's profile yourself; ask your platform operator to move it to one with higher limits.
• Checked on growth only: a quota is checked when something would add to it (a create, an import, a scale-up, an invitation). What you already have keeps running if your limits are later lowered.

How many organizations you can own is a separate, per-account limit; see Create an organization.

Check your usage

Open IAM & SECURITY > Quotas, or View Quotas on the Dashboard. The page shows the active organization (the one in the header's Organization picker): its quota profile and, for each quota, how much is used out of the limit. Two figures can look odd:

• A limit of 0 means your profile doesn't include that resource: you can't create any.
• Usage above 100% happens when your profile was lowered after the resources were created. They keep working, but you can't add more until usage is back under the limit.

Viewing quotas needs resourcemanager_organization_quota_get on the organization. It can't be granted in a project, so members who hold only project grants can't see the page's figures. Organization administrators can.

What each quota counts

The page lists each quota by the identifier below. Your profile may not include all of them.

Quota	What it counts
projects	Projects, including the one created with the organization
users	Organization members, not counting the owner, plus pending invitations
groups	Groups
infrastructures	Infrastructures
service-configs	Service configs
networks, routers, firewalls, ssh-keys	Networks, routers, firewalls and SSH keys
managed-virtual-machines	Virtual machines
mvms-snapshots	Virtual machine backups
managed-kubernetes-clusters	Kubernetes clusters
managed-kubernetes-nodes	Kubernetes nodes (see below)
postgresql-instances	PostgreSQL instances
clusters	Clusters you register for applications
cluster-fleets	Fleets
cluster-fleet-instances	Applications deployed to a fleet
helm-app-templates, container-app-templates	Helm and container templates you create
helm-apps, container-apps	Deployed Helm and container applications, one per cluster

Most quotas count one unit per object. The exceptions:

• Kubernetes nodes: a cluster with a Standalone control plane counts its control plane VMs; a Hosted control plane counts nothing. Each node pool counts its node count, or its Maximum nodes when Autoscale is on. Changing node size doesn't change the count, and Kubernetes nodes never count as virtual machines.
• Virtual machines and PostgreSQL: one unit per instance, whatever its size. Resizing or upgrading never uses quota. The Kubernetes cluster the platform runs behind a Fully managed PostgreSQL instance doesn't count against your Kubernetes quotas.
• Users: an invitation takes a seat as soon as you send it. Declining, cancelling or expiry gives it back; when it is accepted, the seat passes to the new member.
• Clusters: only clusters you register count, not the ones the platform registers for you from your Kubernetes clusters.
• Applications: one unit per cluster. Deploying a Helm application to a fleet of five clusters uses one cluster-fleet-instances unit and five helm-apps units.
• Imported resources count exactly like ones you create, so importing many existing resources can use up a limit. See Imported Resources.

When you hit a limit

If a create, import or scale-up would go over a limit, it is refused and nothing changes. The message says a quota was exceeded and often links to the Quotas page. It doesn't always name the exact quota: creating a Kubernetes cluster, for example, can be refused by either managed-kubernetes-clusters or managed-kubernetes-nodes, so check both.

Deploying an application to a fleet is the exception: clusters that don't fit wait, use no quota, and deploy by themselves within a few minutes after you free quota or your limit is raised. See Applications.

Free up quota

• Delete what you no longer need. A resource stops counting when its deletion finishes, not when you click Delete.
• Clean up failed resources. A resource in Error still counts until you delete it.
• Detach, if you want to keep the cloud object. Detaching frees the quota, but the object keeps running in your cloud account. See Imported Resources.
• Ask for a bigger profile from your platform operator.

Messages that aren't about your profile

• The provider's quota is exceeded (for example when creating an SSH key): the cloud account behind your infrastructure reached its own limit. Raise it with your cloud provider.
• Organization quota is not configured: no limit was reached; the platform has no quota set up for your organization or for that kind of resource. Ask your platform operator to configure it.
• Creating an organization is refused for quota: you already own as many organizations as your account allows. See Create an organization.