HashiCorp Terraform infrastructure-as-code management just got easier, company officials said, attributing the product's recent development in part to funding from its controversial new Business Source License.
Terraform Stacks, released this week in private preview for all editions of Terraform, simplifies large-scale infrastructure-as-code deployments by allowing multiple Terraform modules and workspaces to be updated at the same time from a centralized interface. In the past, updates had to be made individually for separate modules, which often required cutting and pasting code between interdependent Terraform configurations in the correct order, a complex and error-prone process.
Terraform Stacks will be a part of the free Community Edition that replaces the open source version of the tool under HashiCorp's new Business Source License as well as the Cloud and Enterprise editions, said Meghan Liese, vice president of product marketing at HashiCorp.
"These are major enhancements to the execution engine that are being asked [for] from the community. … That's really what we're focused on now," Liese said. "Our move to BSL is really about the investment that allows us to put these types of innovations back into HashiCorp and toward a future that supports both the community and commercial offerings."
Terraform has been at the epicenter of controversy following the BSL change, with a Linux Foundation project, OpenTofu, now forking the open source code and Terraform Registry as a direct response.
Liese declined to comment on OpenTofu specifically but said Terraform's community remains strong and indicated the company has no intention of backing down from its BSL strategy.
"The move to a community version is really about serving … just under 50,000 members across 173 HashiCorp user groups in 71 countries; 38,000 individuals who are HashiCorp cloud engineers and certified on TerraForm, Vault and Consul; and the ecosystem of over 900 partners," she said. "That's who we listen to."
One Terraform Community Edition user said that while he had yet to test Stacks directly, such a feature has long been on his wish list for the product.
"The one piece that was really missing was the ability to hand [Terraform modules] to a developer and say, 'Go spin up a couple of environments if you want to test new features and then kill them when you're done,'" said Rob Lazzurs, director of technology at Amach Software, a digital transformation consultancy in Dublin, Ireland. "[Stacks is] going to solve that problem for us, I believe."
Amach is listed as a supporter of OpenTofu, but neither the company nor any of its clients have plans to move away from the HashiCorp version, Lazzurs said.
"Our position as a consultancy [is] to recommend HashiCorp by default but also publicly make it clear we will support clients who use the fork," he said. "We expect all of our existing clients to stick with Terraform in both the short- and medium-term."
HashiCorp officially embraces genAI for Terraform tests
HashiCorp Terraform added its first official support for generative AI this week, which will follow the general availability of a test framework for the HashiCorp Configuration Language (HCL) introduced in version 1.6 on Oct 4. That release also added beta support for the test-driven development of Terraform modules, similar to what developers already do for other formal programming languages.
Meghan LieseVice president of product marketing, HashiCorp
Terraform users also now have beta access to Terraform tests created by generative AI, along with generally available generative AI-based error correction for HCL code within Visual Studio Code. Users could already point large language models such as GitHub Copilot at Terraform code in the past, but these updates will improve the accuracy of results, Liese said.
"The team has been working with code generation tools to validate code as it comes into the Terraform workflow, … because a lot of things that are generated may not be accurate," she said. "Sometimes it's not quite right."
Lazzurs already uses GitHub Copilot and said he is eager to try out similar features for Terraform.
"Personally, when I've been writing Terraform code since we turned on GitHub Copilot, it's definitely increased my speed during the day," he said. "It's definitely made my life easier."
One analyst called HashiCorp's choice to focus on tests first "a lowkey foray into genAI."
"One might think they would leap into a full on Terraform Copilot, but instead they went to using genAI for generating Terraform module tests [first]," said Katie Norton, an analyst at IDC. Still, "our data suggests this will be desired and valuable for developers."
IDC's "April 2023 Generative AI Adoption and Attitudes" survey of U.S. developers found that software quality and testing was the application development process that had the most potential to benefit from generative AI tools. Out of 200 respondents, 22.5% said software quality and testing will benefit most, followed by 19.5% that chose security testing and vulnerability management.
Beth Pariseau, senior news writer at TechTarget, is an award-winning veteran of IT journalism. She can be reached at [email protected] or on Twitter @PariseauTT.