Cloud

Redpanda BYOVPC Module Options for AWS

This reference lists the inputs of the Redpanda BYOVPC Terraform module for AWS, version 2.3.0, grouped by purpose.

Use this reference to:

  • Look up the inputs of the Redpanda BYOVPC Terraform module for AWS

  • Find the default value of each input

  • Identify the inputs for deploying into existing VPCs and subnets

For the inputs of later releases, see the Inputs tab of the module on the Terraform Registry.

Deprecated inputs are not listed.

General

Input Type Default Description

region

string

None (required)

The AWS region for the VPC and the cluster resources.

aws_account_id

string

""

The ID of the AWS account for the cluster. If empty, the module uses the account of the identity that runs Terraform.

common_prefix

string

"redpanda"

Text added to the start of the names of resources that support name prefixes.

zones

list(string)

["use2-az1", "use2-az2", "use2-az3"]

The availability zone IDs, such as use2-az1, for the subnets that the module creates. The module assigns the subnets to the zones round-robin. If you list fewer than three zones, the module adds randomly selected zones to reach three. When you provide your own subnets with vpc_id, the module ignores this input.

VPC and subnets

By default, the module creates the VPC and subnets. To deploy into a VPC that you manage, set vpc_id and the subnet ID inputs instead of the CIDR inputs. See Deploy with pre-existing VPC and subnets.

Input Type Default Description

vpc_cidr_block

string

"10.0.0.0/16"

The CIDR block of the VPC that the module creates. Ignored when you set vpc_id.

private_subnet_cidrs

list(string)

["10.0.0.0/24", "10.0.2.0/24", "10.0.4.0/24", "10.0.6.0/24", "10.0.8.0/24", "10.0.10.0/24"]

The CIDR blocks of the private subnets. The module creates one private subnet for each CIDR. Ignored when you set private_subnet_ids.

public_subnet_cidrs

list(string)

[]

The CIDR blocks of the public subnets. The module creates one public subnet for each CIDR. Cannot be combined with public_subnet_ids.

vpc_id

string

""

The ID of an existing VPC. When set, the module does not create a VPC. Also set private_subnet_ids. Otherwise, the module creates private subnets in that VPC from private_subnet_cidrs.

private_subnet_ids

list(string)

[]

The IDs of existing private subnets. Provide either private_subnet_cidrs or private_subnet_ids.

public_subnet_ids

list(string)

[]

The IDs of existing public subnets in a VPC that you manage. Required for a dual listener cluster, which places brokers and the internet-facing seed load balancer in public subnets. Each subnet must have a route to an internet gateway and the tag kubernetes.io/role/elb = 1. Cannot be combined with public_subnet_cidrs.

network_exclude_zone_ids

list(string)

[]

Availability zone IDs that the module must not select when it adds zones to reach three. Applies only when the module creates the VPC.

create_internet_gateway

bool

false

When true, the module creates an internet gateway and a route to it for inbound traffic from the Redpanda control plane, even in a VPC that you manage. The module always creates an internet gateway for a VPC that it creates.

public_subnet_map_public_ip_on_launch

bool

true

Sets map_public_ip_on_launch on the public subnets that the module creates. Set to false if a service control policy denies ec2:ModifySubnetAttribute, or if you don’t want instances to receive public IP addresses automatically.

create_s3_gateway_endpoint

bool

true

When true, the module creates an S3 gateway VPC endpoint. Set to false if the VPC already has an S3 gateway endpoint on its private route tables.

Features

These inputs grant the extra permissions, and create the extra resources, that optional cluster features require.

Input Type Default Description

enable_private_link

bool

false

When true, grants the permissions that AWS PrivateLink requires. Also set aws_private_link on the redpanda_cluster resource.

enable_redpanda_connect

bool

true

When true, grants the permissions that Redpanda Connect requires.

enable_redpanda_sql

bool

false

When true, grants the permissions and creates the resources that Redpanda SQL requires. See Enable Redpanda SQL on a BYOVPC Cluster on AWS.

enable_glue_iceberg_catalog

bool

false

When true, allows access to the AWS Glue Data Catalog for the Iceberg REST catalog integration. The module adds the Glue actions to the agent permissions boundary and creates a glue-iceberg policy, both scoped to the Glue resources in this account and region. With enable_redpanda_sql, the module also creates a read-only glue-iceberg-ro policy for Redpanda SQL. After you create the cluster, attach the policy in the glue_iceberg_policy_arn output to the broker role that the agent creates, and the policy in the glue_iceberg_read_only_policy_arn output to the Redpanda SQL role. See the module README.

glue_iceberg_namespaces

list(string)

["redpanda"]

The AWS Glue databases that hold the cluster’s Iceberg topic tables. The read-only Redpanda SQL policy allows access to these databases only. Set this to the value of the iceberg_default_catalog_namespace cluster property if you changed it. Must list at least one database.

enable_public_private_connections

bool

false

When true, the VPC can host a dual listener cluster with both public and private connections. Adds an ingress rule from 0.0.0.0/0 for the public broker ports (30042-30044) to the Redpanda node security group. The private broker ports (30092-30094) stay restricted to private ranges. Requires a public subnet in every availability zone that has a private subnet.

source_cluster_bucket_names

set(string)

[]

The names of the Tiered Storage buckets of the clusters that this cluster reads from as a remote read replica.

reader_cluster_id

string

""

The ID of this Redpanda cluster. Required when you set source_cluster_bucket_names.

IAM

Input Type Default Description

condition_tags

map(string)

{ "redpanda-managed" = "true" }

Tags that restrict the IAM permissions that the module grants. The Redpanda agent can change IAM policies and roles only when they carry all of these tags. Must not be empty.

Redpanda applies redpanda-managed: true to every resource it creates, so don’t set that tag on the redpanda_cluster resource. The Redpanda Terraform provider, version 2.6.0 and later, rejects cluster tag keys that start with redpanda-.

If several clusters share an AWS account, add a tag with the same key and a different value for each cluster, so that one cluster’s agent can’t change another cluster’s IAM resources. Use a key that doesn’t start with redpanda-, and set the same tag in tags on the redpanda_cluster resource. Set these tags when you create the cluster, and don’t add, remove, or change them afterward, here or on the cluster. The agent can’t re-tag the IAM resources that it already created, so it could no longer update or delete them.

create_eks_nodegroup_service_linked_role

bool

true

When true, makes sure that the EKS node group service-linked role exists in the account. New AWS accounts might not have this role. Set to false if you manage the role yourself.

create_rpk_user

bool

false

When true, creates a policy with the minimum permissions for the user who creates the cluster, for testing those permissions. Not for production use.

Storage

Input Type Default Description

force_destroy_cloud_storage

bool

false

When true, terraform destroy deletes the Tiered Storage bucket even if it contains objects. Keep false outside of testing.

force_destroy_management_bucket

bool

true

When true, terraform destroy deletes the management bucket even if it contains versioned Terraform state objects. Set to false in production so that terraform destroy fails instead of deleting the state.

create_rpsql_public_access_block

bool

true

When true and enable_redpanda_sql is true, the module manages a bucket-level setting that blocks all public access to the Redpanda SQL storage bucket. Set to false only when your organization denies the s3:PutBucketPublicAccessBlock action, for example with a service control policy. Before you set it to false on an existing deployment, see the module README.

Tags

Input Type Default Description

default_tags

map(string)

{ }

Tags applied to every resource that the module creates.