We’ve released Terraform 0.15.0-rc1, which marks the end of the prerelease testing period for Terraform v0.15 provided no critical issues are discovered…
During this period, we’d be very grateful if you could help test the release for any unusual behavior.
We do not recommend using the RC in production. While this release has seen some alpha & beta testing, this will be the first exposure of some code to use-cases the Terraform team didn’t anticipate during initial testing. There may well be bugs lurking which we’ll aim to address during the beta period.
Terraform 0.15 is a largely a housekeeping release to prepare for Terraform 1.0. A number of deprecations have been completed with removals and/or errors. Please see the changelog to review how these changes might affect your workflows.
0.15 delivers a number of general quality-of-life improvements including:
- Improved virtual terminal & utf-8 support across all platforms, including Windows.
- Windows support now includes better HCL diagnostic messages.
- We’ve also touched logging, adding structured logging levels beyond “TRACE” for both the CLI and providers (which will themselves need to be updated to support these levels.)
- Values in files for undeclared variables (ex.
tfvars) are no longer deprecated, but will continue to produce a warning.
nonsensitivefunctions allow module authors to explicitly override Terraform’s default infererence of value sensitivity for situations where it’s too conservative or not conservative enough.
-lockfile=readonlyflag, which suppresses writing changes to the dependency lock file.
- Provider-defined sensitive attributes redaction is no longer experimental, but default behavior
One key breaking change is an improvement to modules, where provider aliases can be defined using the
configuration_aliases argument within the
required_providers block. Empty provider configuration blocks should be removed. Read more these changes here.
The draft upgrade guide is here.
We’ve locked this forum topic because we’d like to keep the feedback discussion attached to individual functions (or areas) , so that the discussions are easier to follow. However, if you have any general feedback that isn’t covered by one of these topics we always welcome that via issues in our GitHub repository.