`TF_LOG_PROVIDER_{PROVIDER_NAME}` does not work as documented

[kota65535]

Terraform Version

Terraform v1.13.0
on darwin_arm64

Terraform Configuration Files

The both provider uses tflog package for logging, which is required for using TF_LOG_PROVIDER_{PROVIDER_NAME}.

terraform {
  required_providers {
    aws = {
      source  = "hashicorp/aws"
      version = "~> 6.0"
    }
    google = {
      source  = "hashicorp/google"
      version = "~> 6.0"
    }
  }
}

provider "aws" {
  region = "us-west-2"
}

provider "google" {
  project = "my-gcp-project"
  region  = "us-central1"
}

Debug Output

% TF_LOG_PROVIDER_AWS=debug terraform plan

No changes. Your infrastructure matches the configuration.

Terraform has compared your real infrastructure against your configuration and found no differences, so no changes are needed.

Expected Behavior

Output logs from AWS Provider at the DEBUG level, and no output logs from any other providers

Actual Behavior

No logs from all providers

Steps to Reproduce

% terraform init
% TF_LOG_PROVIDER_AWS=debug terraform plan

Additional Context

According to the documentation (https://developer.hashicorp.com/terraform/plugin/log/managing#enable-logging), we can control the log level for a specific provider by setting TF_LOG_PROVIDER_{PROVIDER_NAME} environment variable. For example, setting TF_LOG_PROVIDER_AWS=trace is expected to output AWS Provider logs at the TRACE level while suppressing logs from other providers. However, this simply does not work.

I've tested various combinations with TF_LOG_PROVIDER, but TF_LOG_PROVIDER_{PROVIDER_NAME} appears to have no effect on log output:

TF_LOG_PROVIDER	TF_LOG_PROVIDER_XXX	Expected Provider Log Level	Actual Provider Log Level
(not set)	DEBUG	XXX: DEBUG Others: OFF	❌ XXX: OFF ✅ Others: OFF
DEBUG	(not set)	XXX: DEBUG Others: DEBUG	✅ Works as expected
DEBUG	INFO	XXX: INFO Others: DEBUG	❌ XXX: DEBUG ✅ Others: DEBUG
INFO	DEBUG	XXX: DEBUG Others: INFO	❌ XXX: INFO ✅ Others: INFO

From my code investigation, Terraform Core is responsible for determining provider log levels. However, the current implementation only recognizes TF_LOG_PROVIDER and TF_LOG environment variables. It does not consider TF_LOG_PROVIDER_{NAME} to set log levels on a per-provider basis.

References

Original discussion: https://discuss.hashicorp.com/t/tf-log-provider-name-does-not-work/39963

Generative AI / LLM assisted development?

No response

[crw]

Thanks for this report!