Multi-runner v2 example

This example demonstrates the experimental multi-runner v2 interface. Shared defaults are configured with global_config* variables, while each runner lane uses multi_runner_config for its matcher, runner lifecycle, and compute-provider settings.

The example creates three lanes from one deployment:

• Linux ARM64 Amazon Linux runners.
• Ephemeral Linux x64 Amazon Linux runners with job retry enabled.
• Windows x64 Server Core 2022 runners.

The v2 interface keeps provider-owned settings inside the selected provider configuration. For example, VPC and subnet settings are under global_config_compute_provider.aws.ec2, while the per-lane instance types and AMI configuration are under each lane's compute provider block. The optional ami variable can provide per-lane AMI filters and owners, which is useful for test environments with locally registered images.

Configure the GitHub App variables before applying:

terraform init
terraform apply \
  -var='github_app={id="123456",key_base64="..."}'

The github_app value is sensitive and should be supplied through a secure variable source in real deployments rather than committed to configuration.

Requirements

Providers

Modules

Resources

Inputs

map(object({
    filter = optional(map(list(string)), { state = ["available"] })
    owners = optional(list(string), ["amazon"])
    id_ssm_parameter = optional(object({
      arn = string
    }), null)
    kms_key = optional(object({
      arn = string
    }), null)
  }))

object({
    id         = string
    key_base64 = string
  })

Outputs