Terraform-provider-aws: Error creating API Gateway Stage: ConflictException: Stage already exists

[chgangaraju]

I am trying to deploy an API in API Gateway with cloudwatch logs enabled. As given in this documentation, it's creating two stages (dev, prod). If i provide two stage names in aws_api_gateway_deployment, aws_api_gateway_stage which is creating two stages. If i give a same name, It's throwing an error saying Stage already exists

It looks like its not possible to create a single stage with cloudwatch logs enabled using terraform?

resource "aws_api_gateway_deployment" "test-deploy" {
  depends_on = [ /*something goes here*/]

  rest_api_id = "${aws_api_gateway_rest_api.test.id}"
  stage_name  = "test"

  variables = {
    "function" = "${var.lambda_function_name}"
  }
}

resource "aws_api_gateway_stage" "test" {
  stage_name = "${var.stage_name}"
  rest_api_id = "${aws_api_gateway_rest_api.test.id}"
  deployment_id = "${aws_api_gateway_deployment.test-deploy.id}"
}

resource "aws_api_gateway_method_settings" "settings" {
  rest_api_id = "${aws_api_gateway_rest_api.test.id}"
  stage_name  = "${aws_api_gateway_stage.test.stage_name}"
  method_path = "*/*"

  settings {
    metrics_enabled = true
    logging_level = "INFO"
  }
}

Error:

aws_api_gateway_stage.test: Error creating API Gateway Stage: ConflictException: Stage already exists

I can't use aws_api_gateway_method_settings without aws_api_gateway_stage, aws_api_gateway_deployment which creates two different stages.

[chgangaraju]

It seems like a documentation problem and the example given here. aws_api_gateway_deployment creates a stage, but i faced a issue saying stage does not exist when i don't have depends_on statement. Its working with the following code.

resource "aws_api_gateway_deployment" "test-deploy" {
  depends_on = [ /*something goes here*/]

  rest_api_id = "${aws_api_gateway_rest_api.test.id}"
  stage_name  = "${var.stage_name}"

  variables = {
    "function" = "${var.lambda_function_name}"
  }
}

resource "aws_api_gateway_method_settings" "settings" {
  depends_on  = ["aws_api_gateway_deployment.test-deploy"]
  rest_api_id = "${aws_api_gateway_rest_api.test.id}"
  stage_name  = "${var.stage_name}"
  method_path = "*/*"

  settings {
    metrics_enabled = true
    logging_level = "INFO"
  }
}

It would be good if this documentation can be updated with a simple example without complicating/creating multiple stages.

[viswaug]

This is still an issue IMHO. I am running into it too. Not have a stage resource works like above since the deployment resource creates it, but it should still be considered a bug since we are not able to use both the resources in the script which seems like a valid thing to do.

Can we consider opening it back up again?

[SachinMule]

Yes, this needs to be reopened.
I get ConflictException when i use "aws_api_gateway_deployment" and "aws_api_gateway_stage" because i want to set the "cache_cluster_size" on the stage.

[Puneeth-n]

Yea, had this issue few months ago. the trick is to depend on all related methods and integration and to just use the deployment and not the stage. The method_settings should then depend on the deployment.

[serialseb]

This prevents assigning a client certificate to the stage that gets implicitly created by the deployment, or any other property on aws_api_gateway_stage, unless I'm misunderstanding something.

[svyandun]

Has anyone figured a way to make this work?
I too get a conflict when trying to enable cache on a stage.

[fmthoma]

I'm running into this bug, too. I need to explicitly configure a stage to associate it with a given certificate.

The AWS documentation says:

--stage-name (string)

The name of the Stage resource for the Deployment resource to create.

--stage-description (string)

The description of the Stage resource for the Deployment resource to create.

So apparently, AWS itself creates a new stage for each deployment. Indeed, Terraform only calls the AWS API:

    deployment, err := conn.CreateDeployment(&apigateway.CreateDeploymentInput{
        RestApiId:        aws.String(d.Get("rest_api_id").(string)),
        StageName:        aws.String(d.Get("stage_name").(string)),
        Description:      aws.String(d.Get("description").(string)),
        StageDescription: aws.String(d.Get("stage_description").(string)),
        Variables:        aws.StringMap(variables),
        })

However, according to the AWS documentation, it seems that --stage-name is not a required parameter. So I think the easy fix would be to make stage_name and stage_description optional in aws_api_gateway_deployment.

[vlad2]

Hello,

Any solution to this issue?

[austinbutler]

Since this issue is closed, commenting on it won't do much. There's an open issue that should be +1'd and followed instead: https://github.com/terraform-providers/terraform-provider-aws/issues/2918

[mhvelplund]

For posterity, the solution is to set stage_name = "" in the aws_api_gateway_deployment resource.

[hashibot[bot]]

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!