# Cedar policy authoring for cluster size and autoscaling caps > Source-anchored research for Cedar policy authoring for cluster size and autoscaling caps. Run status: COMPLETED. Parent: [Atlas Cost Commitments, Cross-Org Billing and FinOps Tooling](https://llms-explorer.com/tree/atlas-cost-commitments-cross-org-billing-and-finops-tooling/) · 5 facets · 14 facts · page: https://llms-explorer.com/tree/cedar-policy-authoring-for-cluster-size-and-autoscaling-caps/ ## How it works - MongoDB documents this inclusive M30–M60 General-class boundary: forbid(principal, action == ResourcePolicy::Action::"cluster.modify", resource) when { (context.cluster has minGeneralClassInstanceSizeValue && context.cluster.minGeneralClassInstanceSizeValue < 30) || (context.cluster has maxGeneralClassInstanceSizeValue && context.cluster.maxGeneralClassInstanceSizeValue > 60) }; (low confidence) (verified as of 2026-10-18) — [source](https://www.mongodb.com/docs/atlas/atlas-resource-policies/) ## How-to and procedures - Use /orgs/{ORG-ID}/nonCompliantResources to detect existing violations and remediate them; applying a policy does not automatically rewrite pre-existing configurations. The independent Foojay walkthrough demonstrates unchanged but flagged resources. (high confidence) — [source](https://www.mongodb.com/docs/atlas/atlas-resource-policies/) ## Problems, failure modes and limitations - Guard optional attributes before comparing them. Cedar skips policies that error during evaluation; MongoDB’s tier example uses has checks. Whether Atlas adds stricter runtime error handling is not established here. (medium confidence) — [source](https://www.mongodb.com/docs/atlas/atlas-resource-policies/) ## Comparisons and alternatives - The minGeneralClassInstanceSizeValue and maxGeneralClassInstanceSizeValue attributes represent the current tier or the minimum and maximum autoscaling sizes; author against these bounds rather than checking only a desired current tier. (low confidence) (verified as of 2026-10-18) — [source](https://www.mongodb.com/docs/atlas/atlas-resource-policies/) - Cedar skips policies that error during evaluation. Issue #2428 illustrates incorrect string membership using in, corrected with set.contains(value), and maintainers identify it as invalid unvalidated policy rather than an engine bug. This is a general authoring caution, not evidence of an Atlas exploit. (high confidence) — [source](https://docs.cedarpolicy.com/auth/authorization.html) - A practitioner walkthrough reports successful Terraform deployment and blocked noncompliant requests, but does not demonstrate enforcement of a tier or autoscaling cap. Treat its cap discussion as guidance rather than an independent cap test. (low confidence) (verified as of 2025-08-13) — [source](https://foojay.io/today/enforcing-governance-in-mongodb-atlas-with-resource-policies/) ## Facts and statements - Atlas Resource Policies add organization-wide configuration guardrails to RBAC; project exceptions allow narrower application. Cluster tier caps are a documented cost-control use. (high confidence) (verified as of 2026-10-18) — [source](https://www.mongodb.com/docs/atlas/atlas-resource-policies/) - MongoDB explicitly excludes Low CPU and NVMe SSD classes from its sample General class tier restriction; do not describe that sample as a universal instance-size cap. (low confidence) (verified as of 2026-10-18) — [source](https://www.mongodb.com/docs/atlas/atlas-resource-policies/) - Atlas accepts forbid policies, excludes permit statements, and allows role-authorized operations by default. Generic Cedar instead defaults to denial without a matching permit; do not transfer its default directly to Atlas. (high confidence) (verified as of 2026-10-18) — [source](https://www.mongodb.com/docs/atlas/atlas-resource-policies/) - Existing autoscaling bounds can exceed a newly introduced policy and continue scaling beyond it, although manual scaling beyond the policy is blocked. Inspect /orgs/{ORG-ID}/nonCompliantResources and reconcile existing bounds. (low confidence) (verified as of 2026-10-18) — [source](https://www.mongodb.com/docs/atlas/atlas-resource-policies/) - Manage resource policies as code with the MongoDB Atlas Terraform provider; an independently authored Foojay walkthrough embeds Cedar bodies in mongodbatlas_resource_policy, while the organization module supports applying policies to existing organizations. (high confidence) — [source](https://foojay.io/today/enforcing-governance-in-mongodb-atlas-with-resource-policies/) - Upper and lower cluster tier constraints were publicly announced on March 3, 2025 for provisioning, scaling and updates; this dated preview announcement is historical evidence, not a statement of current preview status. (medium confidence) (verified as of 2025-03-03) — [source](https://www.mongodb.com/community/forums/t/now-in-public-preview-two-more-atlas-resource-policies-to-manage-cluster-sizes-and-maintenance-windows/314568) - The min/max General-class context values cover the current tier and autoscaling bounds. The example does not constrain Low CPU or NVMe SSD clusters. (low confidence) (verified as of 2026-10-18) — [source](https://www.mongodb.com/docs/atlas/atlas-resource-policies/) - Terraform can carry Cedar in mongodbatlas_resource_policy policies[].body, with org_id and name. Versioning policies alongside infrastructure enables reviewable, repeatable governance changes. (high confidence) (verified as of 2026-10-18) — [source](https://www.mongodb.com/docs/atlas/atlas-resource-policies/)