High CPU and Slow Responses

[rifelpet]

Language Server Version

0.36.4+Homebrew

Terraform Version

Terraform v1.7.0 on darwin_arm64, via tfenv

Client Version

VSCode 1.97.1 w/ hashicorp.terraform 2.34.3

Behavior

After a while of having my terraform directories open in VSCode I'll notice that the terraform-ls process is consistently using >600% CPU.

When this happens and I save a .tf file in VSCode, the Getting code actions from 'HashiCorp Terraform' notification spins for ~20 seconds and until another Saving 'modules_generated.tf': Running Code Actions and Formatters... notification pops up. Both remain spinning for another 7 or 8 minutes. Even after the notifications have disappeared, the file still isn't saved so running terraform commands in my terminal wont pick up the changes I've made.

Steps to Reproduce

Unfortunately I don't have specific instructions. It happens after I've had various VSCode workspaces open for hours.

Most of my VSCode workspaces contain a single terraform root module though they are subdirectories of a larger monorepo with thousands of terraform root and shared modules.

Here is a screenshot from Activity Monitor showing the sustained CPU usage:

In addition to the logs gist below, I also uploaded base64-encoded profile dumps:

• memory: https://gist.github.com/rifelpet/b1c817ab13c1a069d1675c2bf2747b65
• cpu (large): https://gist.github.com/rifelpet/4c4c129e46a2904aba808d7e903b8a01
  Let me know if you have trouble loading those and I can email them to a hashicorp address.

Terraform Configuration

Project Structure

Gist

https://gist.github.com/rifelpet/a20827b2345aac0fe06ab0a7fabac2e4

References

No response

Community Note

• Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritize this request
• Please do not leave "+1" or other comments that do not add relevant new information or questions, they generate extra noise for issue followers and do not help prioritize the request
• If you are interested in working on this issue or have submitted a pull request, please leave a comment