weekly.tf #19 - 0.13 released

Big releases recently in the Terraform world- 0.13 and AWS provider 3 are both out and read to be used.

HashiCorp has definitely gotten better at making these low-stress (from the outside).

After a few months of beta testing, Terraform 0.13 has been released, featuring improvements to working with modules and distributing providers.

Now that count and for_each are allowed on modules there must be many teams who can simplify a lot of code. Variable validation for modules will make the process of writing robust modules easier.

I have been trying this out for our snowflake provider and have to say that being able to publish providers to the registry is a dramatic improvement over the hand-rolled processes I have seen.

The icing on the cake is integrated, version docs.

A new tier of service is available for Terraform Cloud, including a number of features previously reserved for Enterprise.

In addition to SSO with Okta, job concurrency and an SLA you can monitor activity with Splunk (if you are into that sort of thing).

Here's the thing– those features are all pretty straightforward and probably will convince a few teams they can use Cloud rather than Enterprise. The real difference maker here is the ability to run private agents.

How many teams are running Enterprise because they either need access to some internal resource or to keep their credentials in their networks. Now that private agents are supported, the calculus for cloud vs on-prem is going to change. And running Terraform Enterprise kinda sucks.

Following up from the major 3.0.0 release, 3.1 is a handful of new features and quite a few bug fixes. Still enough here that it is hard to summarize.

As a VS Code user, the work on the language server has been the biggest day-to-day improvement in Terraform in a long time. This release continues that process. I am particularly interested in #256, which fixes a problem I have been running into.