Terraform module which creates VPC resources on AWS.

These types of resources are supported:

• VPC
• Subnet
• Route
• Route table
• Internet Gateway
• Network ACL
• NAT Gateway
• VPN Gateway
• VPC Endpoint:
  • Gateway: S3, DynamoDB
  • Interface: EC2, SSM, EC2 Messages, SSM Messages, SQS, ECR API, ECR DKR, API Gateway, KMS, ECS, ECS Agent, ECS Telemetry, SNS, CloudWatch(Monitoring, Logs, Events), Elastic Load Balancing, CloudTrail
• RDS DB Subnet Group
• ElastiCache Subnet Group
• Redshift Subnet Group
• DHCP Options Set
• Default VPC
• Default Network ACL

Sponsored by Cloudcraft - the best way to draw AWS diagrams

Terraform 0.12. Pin module version to ~> v2.0. Submit pull-requests to master branch.

Terraform 0.11. Pin module version to ~> v1.0. Submit pull-requests to terraform011 branch.

module "vpc" {
  source = "terraform-aws-modules/vpc/aws"

  name = "my-vpc"
  cidr = "10.0.0.0/16"

  azs             = ["eu-west-1a", "eu-west-1b", "eu-west-1c"]
  private_subnets = ["10.0.1.0/24", "10.0.2.0/24", "10.0.3.0/24"]
  public_subnets  = ["10.0.101.0/24", "10.0.102.0/24", "10.0.103.0/24"]

  enable_nat_gateway = true
  enable_vpn_gateway = true

  tags = {
    Terraform = "true"
    Environment = "dev"
  }
}

By default this module will provision new Elastic IPs for the VPC's NAT Gateways. This means that when creating a new VPC, new IPs are allocated, and when that VPC is destroyed those IPs are released. Sometimes it is handy to keep the same IPs even after the VPC is destroyed and re-created. To that end, it is possible to assign existing IPs to the NAT Gateways. This prevents the destruction of the VPC from releasing those IPs, while making it possible that a re-created VPC uses the same IPs.

To achieve this, allocate the IPs outside the VPC module declaration.

resource "aws_eip" "nat" {
  count = 3

  vpc = true
}

Then, pass the allocated IPs as a parameter to this module.

module "vpc" {
  source = "terraform-aws-modules/vpc/aws"

  # The rest of arguments are omitted for brevity

  enable_nat_gateway  = true
  single_nat_gateway  = false
  reuse_nat_ips       = true                    # <= Skip creation of EIPs for the NAT Gateways
  external_nat_ip_ids = "${aws_eip.nat.*.id}"   # <= IPs specified here as input to the module
}

Note that in the example we allocate 3 IPs because we will be provisioning 3 NAT Gateways (due to single_nat_gateway = false and having 3 subnets). If, on the other hand, single_nat_gateway = true, then aws_eip.nat would only need to allocate 1 IP. Passing the IPs into the module is done by setting two variables reuse_nat_ips = true and external_nat_ip_ids = "${aws_eip.nat.*.id}".

This module supports three scenarios for creating NAT gateways. Each will be explained in further detail in the corresponding sections.

• One NAT Gateway per subnet (default behavior)
  • enable_nat_gateway = true
  • single_nat_gateway = false
  • one_nat_gateway_per_az = false
• Single NAT Gateway
  • enable_nat_gateway = true
  • single_nat_gateway = true
  • one_nat_gateway_per_az = false
• One NAT Gateway per availability zone
  • enable_nat_gateway = true
  • single_nat_gateway = false
  • one_nat_gateway_per_az = true

If both single_nat_gateway and one_nat_gateway_per_az are set to true, then single_nat_gateway takes precedence.

By default, the module will determine the number of NAT Gateways to create based on the the max() of the private subnet lists (database_subnets, elasticache_subnets, private_subnets, and redshift_subnets). The module does not take into account the number of intra_subnets, since the latter are designed to have no Internet access via NAT Gateway. For example, if your configuration looks like the following:

database_subnets    = ["10.0.21.0/24", "10.0.22.0/24"]
elasticache_subnets = ["10.0.31.0/24", "10.0.32.0/24"]
private_subnets     = ["10.0.1.0/24", "10.0.2.0/24", "10.0.3.0/24", "10.0.4.0/24", "10.0.5.0/24"]
redshift_subnets    = ["10.0.41.0/24", "10.0.42.0/24"]
intra_subnets       = ["10.0.51.0/24", "10.0.52.0/24", "10.0.53.0/24"]

Then 5 NAT Gateways will be created since 5 private subnet CIDR blocks were specified.

If single_nat_gateway = true, then all private subnets will route their Internet traffic through this single NAT gateway. The NAT gateway will be placed in the first public subnet in your public_subnets block.

If one_nat_gateway_per_az = true and single_nat_gateway = false, then the module will place one NAT gateway in each availability zone you specify in var.azs. There are some requirements around using this feature flag:

• The variable var.azs must be specified.
• The number of public subnet CIDR blocks specified in public_subnets must be greater than or equal to the number of availability zones specified in var.azs. This is to ensure that each NAT Gateway has a dedicated public subnet to deploy to.

By default, if NAT Gateways are enabled, private subnets will be configured with routes for Internet traffic that point at the NAT Gateways configured by use of the above options.

If you need private subnets that should have no Internet routing (in the sense of RFC1918 Category 1 subnets), intra_subnets should be specified. An example use case is configuration of AWS Lambda functions within a VPC, where AWS Lambda functions only need to pass traffic to internal resources or VPC endpoints for AWS services.

Since AWS Lambda functions allocate Elastic Network Interfaces in proportion to the traffic received (read more), it can be useful to allocate a large private subnet for such allocations, while keeping the traffic they generate entirely internal to the VPC.

You can add additional tags with intra_subnet_tags as with other subnet types.

Sometimes you need to have a way to create VPC resources conditionally but Terraform does not allow to use count inside module block, so the solution is to specify argument create_vpc.

# This VPC will not be created
module "vpc" {
  source = "terraform-aws-modules/vpc/aws"

  create_vpc = false
  # ... omitted
}

Sometimes it is handy to have public access to RDS instances (it is not recommended for production) by specifying these arguments:

  create_database_subnet_group           = true
  create_database_subnet_route_table     = true
  create_database_internet_gateway_route = true

  enable_dns_hostnames = true
  enable_dns_support   = true

This module can manage network ACL and rules. Once VPC is created, AWS creates the default network ACL, which can be controlled using this module (manage_default_network_acl = true).

Also, each type of subnet may have its own network ACL with custom rules per subnet. Eg, set public_dedicated_network_acl = true to use dedicated network ACL for the public subnets; set values of public_inbound_acl_rules and public_outbound_acl_rules to specify all the NACL rules you need to have on public subnets (see variables.tf for default values and structures).

By default, all subnets are associated with the default network ACL.

Sometimes it is handy to have public access to Redshift clusters (for example if you need to access it by Kinesis - VPC endpoint for Kinesis is not yet supported by Redshift) by specifying these arguments:

  enable_public_redshift = true  # <= By default Redshift subnets will be associated with the private route table

• Simple VPC
• Complete VPC
• Manage Default VPC
• Network ACL
• Few tests and edge cases examples: #46, #44, #108

This module has been packaged with awspec tests through test kitchen. To run them:

1. Install rvm and the ruby version specified in the Gemfile.
2. Install bundler and the gems from our Gemfile:

gem install bundler; bundle install

3. Test using bundle exec kitchen test from the root of the repo.

Module is maintained by Anton Babenko with help from these awesome contributors.

Apache 2 Licensed. See LICENSE for full details.