HashiCorp has launched a new way to enforce rules on Terraform, called tfpolicy, and the older tool many organisations pay for, Sentinel, is not where the new features are going. Nothing breaks today. But if your governance runs on Sentinel, there is now a shelf-life question to put in this quarter's planning.
No vendor says the product you rely on is legacy. A new thing ships "built for" the platform, the old thing is "still supported", and a migration tool appears near the bottom of the post. On 16 July, HashiCorp's tfpolicy announcement managed all three in one page. tfpolicy is in public beta in HCP Terraform now.
What it is, in plain terms
Policy as code means writing your rules as checks that run automatically instead of as wiki pages: every storage account blocks public access, and nobody deploys a module the platform team hasn't approved. The check runs on every change, before the change happens.
On HashiCorp's platform that has meant Sentinel since 2017, or Open Policy Agent. Both work, and both carry the same tax: the rules are written in a language your team uses for nothing else. In practice one or two people can write policy, every request queues behind them, and governance moves at the speed of a side project.
tfpolicy removes the tax. Policies are written in HCL, the same language as Terraform itself, and kept in version control like everything else, so any Terraform engineer you employ can read them (there is an example in the technical section at the end). Per the announcement, it can also follow relationships between resources, pull in outside data, block unapproved providers and modules before they are downloaded (a supply-chain control as much as a configuration check), and re-check infrastructure after deployment.
How to read the announcement
Sentinel is not deprecated. HashiCorp confirmed to trade press that Sentinel and OPA remain supported, and nothing sets an end date.
The new capabilities land in tfpolicy and not in Sentinel, though, and the same release includes an AI agent skill whose listed jobs are writing and testing tfpolicy files "or converting existing Sentinel policies". Vendors do not usually fund conversion tooling for a product they plan to keep investing in. That is my judgement, not a stated fact. tfpolicy is still a beta, so try it in a sandbox before you move anything.
| Your situation | What to do this quarter |
|---|---|
| You run Sentinel | Count the policies, find out who can change them, and try the converter on a copy |
| You never adopted policy as code | If staffing a second language was the reason, look again |
| You don't use HCP Terraform | Nothing here yet; check that Azure Policy is assigned across the estate |
If you own a Sentinel estate
Organisations bought Terraform Enterprise and the upper HCP Terraform tiers partly for Sentinel. The risk is years of platform-team work sitting in a policy library that fewer people can read, while the vendor's attention is elsewhere.
Ask your platform lead four things:
- How many Sentinel policies do we run, and how many are mandatory rather than advisory? The mandatory ones have to behave identically after any migration.
- Who can change them? If the answer is one name, that was a risk before July too.
- What does the converter produce from a copy of our library? An afternoon with a copy gives you a first estimate of the migration effort.
- What would we want GA to look like before we move? Write the criteria down now: which paid tier it lands in, and how much of the library converts cleanly.
Until you have those answers, pause net-new Sentinel work unless something regulatory forces it. Some of the new rules may belong in Azure Policy anyway.
If you never adopted policy as code
Plenty of estates run Terraform with no policy layer, and the usual reason is that nobody wanted to staff a second language for a part-time job. With tfpolicy that reason is gone.
First check where enforcement runs. The tfpolicy CLI validates and tests policies locally, but enforcement on every run is a feature of HCP Terraform, the paid managed platform. IBM has owned HashiCorp since February 2025, pricing is per managed resource, and the free tier changed shape this spring (the legacy free plan retired on 31 March 2026; the current one covers 500 managed resources). So using tfpolicy to enforce anything means paying for the platform it runs on.
The layer a plan check cannot see
tfpolicy evaluates the plan, so it evaluates what goes through the pipeline. It has no view of the firewall rule someone opened in the portal on a Friday afternoon, because that change never generates a plan.
On Azure the runtime layer is Azure Policy, which you already own: a built-in catalogue of roughly 5,000 rules that costs nothing to run and that most estates leave switched off. With only the gate you miss portal changes; with only the audit you find mistakes after they deploy.
tfpolicy also ships platform-first, with the open CLI getting only the local-testing version. At renewal, ask whether that is how Terraform features will arrive from now on.
For your platform team: what a tfpolicy policy looks like
Policies live in files ending .policy.hcl. From the reference documentation, a complete policy that stops anyone using an outdated version of an approved module:
module_policy "git::github.com/org/terraform-aws-vpc" "example" {
enforce {
condition = core::semverconstraint(meta.version, ">= 2.0.0")
error_message = "The official VPC module must be at version 2.0.0 or higher."
info_message = "VPC module version ${meta.version}."
}
}