Terraform Cloud vs Enterprise vs Self-Hosted

HCP Terraform, Terraform Enterprise, and self-hosted backends compared for 2026 - RUM pricing, migration cost, and when self-hosting still wins.

Frequently Asked Questions

Why did a team's HCP Terraform bill jump 40% with no new environments added?

Because Resources Under Management (RUM) pricing bills on total managed resources, not on how often you run terraform apply — a module added months earlier that manages thousands of granular resources (security group rules, IAM bindings, pod annotations) can inflate the bill quietly long after it was deployed.

Is the old unlimited-resource HCP Terraform free tier still available?

No — the legacy free tier under the old user-based model ended March 31, 2026. Existing organizations were automatically moved to the new 500-resource Free tier, a meaningfully smaller allowance for any team with real production infrastructure.

Is there a hard deadline for teams running Terraform Enterprise on the Replicated installation method?

Yes — HashiCorp ended support for Replicated installs on April 1, 2026, and teams must migrate to Docker, Kubernetes, Podman, or Nomad. Existing air-gapped Replicated installs get a longer runway to December 31, 2027, but no new installs or upgrades are possible on Replicated.

Why do Kubernetes modules often produce far more Terraform resources than teams expect?

Because a single logical unit — like a security group with 9 rules — can count as 10 separate resources in state, and nested modules combined with for_each over long lists routinely produce 3-5x more actual resources than a rough "50 EC2 instances" style headcount estimate suggests.

When does a bare S3-plus-DynamoDB backend make more sense than HCP Terraform?

When your team has strong existing CI/CD investment, doesn't need built-in policy-as-code or drift detection, and your resource count is growing faster than your run frequency — RUM pricing penalizes that exact profile hardest, while a bare backend or a per-run third-party platform doesn't scale cost with resource count the same way.

Discussion0