reserved_concurrent_executions option of aws_lambda_function gives error message

[maanenh]

The reserved_concurrent_executions option of aws_lambda_function does not work.

Terraform Version

• Terraform v0.11.3
• provider.archive v1.0.0
• provider.aws v1.10.0

Affected Resource(s)

aws_lambda_function

Terraform Configuration Files

data "archive_file" "my_lambda" {
  type = "zip"
  source_file = "./myLambda.js"
  output_path = "myLambda.zip"
}

resource "aws_lambda_function" "my_lambda" {
  filename         = "${data.archive_file.my_lambda.output_path}"
  function_name    = "concurrentLambdaIssue"
  role             = "${var.role_arn}"
  handler          = "myLambda.handler"
  runtime          = "nodejs6.10"
  source_code_hash = "${base64sha256(file("${data.archive_file.my_lambda.output_path}"))}"
  
  reserved_concurrent_executions = 2
}

Debug Output

First execution:

• aws_lambda_function.my_lambda: Error setting concurrency for Lambda concurrentLambdaIssue: ResourceNotFoundException: An error occurred and the request cannot be processed.
  status code: 404, request id: xx

When you run terraform apply again the messages is:

• aws_lambda_function.my_lambda: Error creating Lambda function: ResourceConflictException: Function already exist: concurrentLambdaIssue
  status code: 409, request id: xx

Panic Output

n/a

Expected Behavior

Function created with reserved_concurrent_executions set.

Actual Behavior

See error messages.

Steps to Reproduce

1. terraform apply

Important Factoids

• On first execution the message is: ResourceNotFoundException. When you execute terraform apply again the messages is: Function already exist.
• It seems the function is created during the first execution, but the state-files is not updated with this information.
• When you look with the AWS console, the function is indeed created, but without reserved_concurrent_executions

References

[bflad]

Hi @maanenh 👋 thank you very much for this bug report and sorry you're running into trouble here.

We actually acceptance test this regularly without issue but looks like there could be some latent eventual consistency within the AWS Lambda service. There's two fixes we can apply here to help the situation:

• Calling d.SetId() before we call PutFunctionConcurrency so any errors returned by that won't prevent Terraform from seeing the existing resource (it'll just show a plan to update the concurrency again if it errored on creation)
• Retrying for up to a minute on the ResourceNotFound error with PutFunctionConcurrency which should catch eventual consistency issues in most cases

I have a PR ready to go for this, which I'm acceptance testing now!

[bflad]

Both the above fixes have been implemented into master and will be released with v1.11.0 of the AWS provider, likely at the end of this week.

[maanenh]

Thanks for fixing this issue.

[bflad]

This has been released in version 1.11.0 of the AWS provider. Please see the Terraform documentation on provider versioning or reach out if you need any assistance upgrading.

[ghost]

I'm going to lock this issue because it has been closed for 30 days ⏳. This helps our maintainers find and focus on the active issues.

If you feel this issue should be reopened, we encourage creating a new issue linking back to this one for added context. Thanks!