Infracost v0.10.33 being Killed

[mapperCoderZ]

Hello.

I have an infracost command failing on v0.10.33 but not failing with v0.10.31.
We try to validate our terraform base cost on last version of terraform inside a docker container ARM64 and AMD64. We use azure rm provider.

here is my pipeline logs:

rm -f /tmp/infracost-base.json && infracost breakdown --path /wspace/az-tf/infra/zones/azure/mkp-int1 --terraform-var-file /wspace/az-tf/mkp-hpr/azure/mkp-int1/terraform.tfvars
2024-02-15T16:37:32Z INF Evaluating Terraform directory at /builds/infra-cloud/april-gitops-env-mkp-hpr/infra/zones/azure/mkp-int1
2024-02-15T16:37:32Z INF Starting: Downloading Terraform modules
...
2024-02-15T16:39:04Z WRN Using eastus for resource module.application_gateway.azurerm_private_dns_a_record.private_dns_zone_record_application_digital["asp-conn"] as its 'location' property could not be found.
❌ Command "infracost breakdown --path /builds/infra-cloud/april-gitops-env-mkp-hpr/infra/zones/azure/mkp-int1 --terraform-var-file /builds/infra-cloud/april-gitops-env-mkp-hpr/mkp-hpr/azure/mkp-int1/terraform.tfvars --format=json --out-file /tmp/infracost-base.json > /dev/null" failed

Locally, I have this error without other message:
Killed

Can you have a look ?

[vdmgolub]

Hi! Thanks for creating the issue.

Unfortunately, it's pretty hard to debug such issues without the code. Could you please try rerunning it with --log-level=debug or the INFRACOST_LOG_LEVEL=debug env var to see if it provides more info? You can also join our community Slack and find us on the #help channel if you're not comfortable sharing the logs here.

[mapperCoderZ]

t's really painfull to debug such things in our enterprise environment. The debug logs are very long and I need to relaunch. Code base is large, the default stderr output should be improved

[mapperCoderZ]

Maybe the process takes too much memory ?

[vdmgolub]

This is possible if you have a huge repo. Maybe you'll be able to run this command tree -P '*.hcl|*.tf|*.tfvars|*.tfvars.json|*.json' --prune and send us the tree output, so we can see the repo structure?

[vdmgolub]

@mapperCoderZ If it makes things easier, we could screenshare. Please drop a message to [email protected] and we'll send a Calendly link to set it up. Thanks!

[alikhajeh1]

@mapperCoderZ just adding to what @vdmgolub mentioned, I think you're already an Infracost Cloud customer - email us and we'll arrange a quick debugging session and get you to success 🚀

[matthieupetite]

Hello I have reproduce the case and indeed revert to v31 is a good workaround and we also revert the infracost vscode extension to v20. I try to send to you the whole debug script

[matthieupetite]

here is the command I launched

here is the content of my configuration file:

`version: 0.1
projects:

• path: /wspace/az-tf/infra/zones/azure/mkp-int1
  name: mkp-int1
  usage_file: /wspace/az-tf/mkp-hpr/azure/mkp-int1/infracost-usage.yml
  terraform_var_files:
  • /wspace/az-tf/mkp-hpr/azure/mkp-int1//terraform.tfvars
• path: /wspace/az-tf/infra/zones/azure/mkp-int2
  name: mkp-int2
  usage_file: /wspace/az-tf/mkp-hpr/azure/mkp-int2/infracost-usage.yml
  terraform_var_files:
  • /wspace/az-tf/mkp-hpr/azure/mkp-int2/terraform.tfvars
• path: /wspace/az-tf/infra/zones/azure/mkp-rec1
  name: mkp-rec1
  usage_file: /wspace/az-tf/mkp-hpr/azure/mkp-rec1/infracost-usage.yml
  terraform_var_files:
  • /wspace/az-tf/mkp-hpr/azure/mkp-rec1/terraform.tfvars
• path: /wspace/az-tf/infra/zones/azure/mkp-rec2
  name: mkp-rec2
  usage_file: /wspace/az-tf/mkp-hpr/azure/mkp-rec2/infracost-usage.yml
  terraform_var_files:
  • /wspace/az-tf/mkp-hpr/azure/mkp-rec2/terraform.tfvars
• path: /wspace/az-tf/infra/zones/azure/mkp-ppr
  name: mkp-ppr
  usage_file: /wspace/az-tf/mkp-ppr/azure/mkp-ppr/infracost-usage.yml
  terraform_var_files:
  • /wspace/az-tf/mkp-ppr/azure/mkp-ppr/terraform.tfvars
• path: /wspace/az-tf/infra/zones/azure/mkp-pro
  name: mkp-pro
  usage_file: /wspace/az-tf/mkp-pro/azure/mkp-pro/infracost-usage.yml
  terraform_var_files:
  • /wspace/az-tf/mkp-pro/azure/mkp-pro/terraform.tfvars
    `

you will find the log file attached in this post
infracost.log.gz

[matthieupetite]

I noticed the same sort of issue in the vscode tasks

[hugorut]

Thanks @matthieupetite, what happens if you run the infracost breakdown on v31 with the following env variable: INFRACOST_GRAPH_EVALUATOR=true? Do you still see killed?

[matthieupetite]

I don't have any issue with that configuration the command runs successfully

[hugorut]

Thanks @matthieupetite, ok a follow up with two questions:

1. if you modify your config file and add the key skip_autodetect: true to each defined project, then rerun the breakdown (without the INFRACOST_GRAPH_EVALUATOR env), does this work?
2. What happens if you run infracost breakdown --path /wspace/az-tf (again without the INFRACOST_GRAPH_EVALUATOR env)

🙏

[matthieupetite]

Hi Hugo thanks for your support. I will be able to make that test Monday and I will share the result here

[matthieupetite]

Hi Hugo, I have test your procedure and with the new version of Infracost and It is failling now

[matthieupetite]

Hi all, any news on that subject ?

[aliscott]

@matthieupetite we've not got to the bottom of it yet, but it's on our current sprint to investigate further so I will reach out to you sometime in the next 2 weeks.

[matthieupetite]

@matthieupetite we've not got to the bottom of it yet, but it's on our current sprint to investigate further so I will reach out to you sometime in the next 2 weeks.

thanks a lot