Writing terraform output to JSON file

Terraspace
1

Hello BoltOps Community,

I managed to figure out how to use a helper function/method/module to read e.g. a JSON file for input and generate terraform code using ERB templating. I am wondering if we are able to store the output from terraform back into the JSON file?

Here’s an example use case…

Let’s say I have this data populated in config/vpc_config.json

{
    "network": {
        "vpcs": {
            "main": {
                "name": "main",
                "vpc_cidr": "10.98.20.0/22",
                "subnets": {
                    "public": {
                        "az1": {
                            "subnet_type": "external",
                            "subnet_name": "public",
                            "subnet_az"  : "1",
                            "subnet_cidr": "10.98.20.0/26"
                        },
                        "az2": {
                            "subnet_type": "external",
                            "subnet_name": "public",
                            "subnet_az"  : "2",
                            "subnet_cidr": "10.98.20.64/26"
                        }
                    },
                    "private": {
                        "az1": {
                            "subnet_type": "internal",
                            "subnet_name": "private",
                            "subnet_az"  : "1",
                            "subnet_cidr": "10.98.21.0/26"
                        },
                        "az2": {
                            "subnet_type": "internal",
                            "subnet_name": "private",
                            "subnet_az"  : "2",
                            "subnet_cidr": "10.98.21.64/26"
                        }
                    },
                    "database": {
                        "az1": {
                            "subnet_type": "internal",
                            "subnet_name": "database",
                            "subnet_az"  : "1",
                            "subnet_cidr": "10.98.22.0/26"
                        },
                        "az2": {
                            "subnet_type": "internal",
                            "subnet_name": "database",
                            "subnet_az"  : "2",
                            "subnet_cidr": "10.98.22.64/26"
                        }
                    }
                }
            }
        }
    }
}

In my helper file, I have

require 'json'

module Terraspace::Project::CustomHelpers
  def vpc_config_file
    "config/vpc_config.json"
  end

  def vpc_data
    file = File.read(vpc_config_file)
    data_hash = JSON.parse(file)
  end

end

In my vpc.tf file, I am creating a VPC. e.g.

resource "aws_vpc" "vpc" {
  cidr_block = <%= vpc_data['network']['vpcs']['main']['vpc_cidr'] %>
}

Now, I would like to store the value of the vpc_id (which is only known after the VPC is created) in config/vpc_config.json.

I know I can update the config/vpc_config.json by using a method like below

  def vpc_id_writeback(vpc, vpcid)
    data_hash = read_json(vpc_config_file)
    data_hash['network']['vpcs'][vpc]['vpc_id']=vpcid
    File.write(vpc_config_file, JSON.pretty_generate(data_hash))
  end

But I guess, where I’m coming from is… The ERB templating in the .tf files are applied during the generation of .tf files (when we do terraspace build)… Whereas, the value of vpc_id is only known after we run terraspace up…

I understand that we would also be able to query the vpc_id through the use of terraform’s data "aws_vpc" (without having to store the information in the JSON file), but am still interested to know if we’d be able to store the information back into the JSON file - just thinking that if we have to deal with multiple VPCs/looping, it might be easier to loop through using ERB template over terraform’s loop

Many thanks,
James

2

Hello again,

I’ve been poking around the terraspace documentation again and I realized that I’d be able to do what I want to do without needing to write back to the JSON file. I.e. I’d be able to reference the vpc_id without needing to use terraform’s data "aws_vpc" (when they are on a different stacks) by using terraspace’s output helper (with examples in the “How to Configure Dependencies” section here) :smiley:

Nevertheless, am interested to hear back on whether it would be possible (and how) to write information back to the JSON file after the resource is created

Cheers,
James