Google Deployment Manager Replacement — when and how should I migrate?

Choose whether to move Google Cloud infrastructure definitions to Infrastructure Manager or to a self-managed Terraform workflow now that Deployment Manager support ended on March 31, 2026.

Google Cloud Infrastructure Manager, unless you need private repo runs, custom policy gates, newer Terraform, or cross-cloud parity; choose self-managed Terraform only then.

Blockers

Who this is for

Candidates

Google Cloud Infrastructure Manager

As of 2026-04-02, Deployment Manager has already shut down, and Google Cloud's official replacement path is Infrastructure Manager. Infra Manager is a managed Terraform execution environment on Google Cloud, and Google documents that it stores state in Cloud Storage and uses Cloud Build to execute provisioning. As of 2026-04-02, its pricing is usage-based through Cloud Build execution and Cloud Storage storage rather than a separate flat control-plane fee. Google also publishes a supported Terraform matrix for Infra Manager; the docs list 1.3.10, 1.4.7, and 1.5.7 as supported, while 1.2.3 reached end of support on 2026-02-08.

When to choose

Use this when you want the lowest-ops path and are comfortable standardizing on Google Cloud's managed Terraform workflow. It is the default choice if your main goal is replacing Deployment Manager quickly without building your own runner, state handling, and preview/apply plumbing.

Tradeoffs

Managed execution, IAM integration, previews, and automatic state handling reduce platform work. The tradeoff is tighter product constraints than a self-managed Terraform workflow, including published service quotas, version gating, and Google Cloud-specific workflow assumptions.

Cautions

Google documents several migration caveats: Infra Manager does not perform data migration or ensure service continuity, revisions can delete or replace resources, rollback is manual by redeploying an earlier Terraform configuration, and DM Convert cannot export some Deployment Manager concepts because Terraform does not support them. Infra Manager also lacks direct support for private Git repositories, and its documented quotas include 1,000 deployments per project per region, 20 mutating requests per minute per project per region, 150 Terraform input variables, and a 950,000-byte API payload limit.

Sources

Self-managed Terraform workflow

As of 2026-04-02, Google states that Deployment Manager users should migrate to Infrastructure Manager or another deployment technology, and Google documents that Infra Manager uses Terraform configurations. This supports using Terraform HCL with your own CI/CD or local execution instead of Infra Manager. As of 2026-04-02, Google does not publish pricing for this self-managed path, but this option is not bound to Infra Manager's documented execution limits or supported-version matrix.

When to choose

Use this when you already have mature CI/CD, need tighter control over execution, or need to avoid Infra Manager-specific constraints. It is the better choice when private repository workflows, custom policy gates, newer Terraform versions, or cross-cloud consistency matter more than having Google manage the execution layer.

Tradeoffs

You keep Terraform portability and can design your own state backend, runners, approvals, and release process. The cost is that you now own orchestration, locking, secrets, drift workflow, and operational reliability instead of delegating them to Google Cloud.

Cautions

This recommendation is partly an inference from Google's official docs: Google confirms Terraform is the target model and explicitly permits an alternative deployment technology, but it does not prescribe your exact self-managed architecture. You still inherit Terraform-side migration caveats from Deployment Manager, including unsupported DM concepts in DM Convert and the fact that some Google Cloud-specific features or integrations might not be available through Terraform providers.

Sources

Facts updated: 2026-04-02
Published: 2026-04-03

Related questions

Try with your AI agent

$ npm install -g pocketlantern
$ pocketlantern init
# Restart Claude Code, Cursor, or your MCP client, then ask:
# "Google Deployment Manager Replacement — when and how should I migrate?"
Missing something? Request coverage