Legacy logging configuration causes incorrect log level setting

[the-nando]

Terraform Version

Terraform v1.5.5
on darwin_arm64

Affected Resource(s)

Please list the resources as a list, for example:

• resource_pki_secret_backend_cert.go
  Probably more.

Terraform Configuration Files

resource "vault_pki_secret_backend_cert" "static" {
  for_each              = local.additional_certificates
  backend               = var.vault_pki_mount
  name                  = var.vault_pki_role
  common_name           = each.value.common_name
  ttl                   = each.value.ttl
  auto_renew            = true
  min_seconds_remaining = ceil(each.value.ttl / 3)
  alt_names             = each.value.alt_names
}

Debug Output

2023-08-14T15:38:53.408+0200 [INFO]  provider.terraform-provider-vault_v3.17.0_x5: 2023/08/14 15:38:53 [WARN] Failed to check if mount "pki_foo" exist, preempting the read operation: timestamp=2023-08-14T15:38:53.407+0200

Expected Behavior

Logs are captured by Terraform under the correct logging level WARN.

Actual Behavior

Due to the prefixed timestamp, Terraform fails to parse the logging level and the warning message is silently swallowed under INFO level. The provider should strip any prefixed date or time stamp before emitting the log or upgrade the logger:

log.SetFlags(log.Flags() &^ (log.Ldate | log.Ltime))

References

• https://developer.hashicorp.com/terraform/plugin/log/writing#legacy-log-troubleshooting