Bulk Import Existing Kestra Resources

Bulk Import Existing Kestra Resources

Terraform 1.14 introduced bulk import with the list block and terraform query. The Kestra provider exposes list resources for flows and namespaces so existing resources can be discovered before generating Terraform configuration.

Requirements

Use Terraform 1.14 or newer and configure the Kestra provider against the Kestra instance that contains the unmanaged resources.

terraform {
  required_version = ">= 1.14.0"

  required_providers {
    kestra = {
      source  = "kestra-io/kestra"
      version = "~> 2.0"
    }
  }
}

Configure the provider with the old Kestra instance:

provider "kestra" {
  url      = var.kestra_url
  username = var.kestra_username
  password = var.kestra_password
}

Query flows

The flow list resource can return flow identities only:

list "kestra_flow" "all_flows" {
  provider = kestra
  limit    = 10000
}

Set include_resource = true when the query should also retrieve the resource state needed by Terraform’s generated import configuration:

list "kestra_flow" "all_flows" {
  provider         = kestra
  limit            = 10000
  include_resource = true
}

The provider paginates the Kestra API internally. A limit larger than the Kestra API page size is handled across multiple API pages.

Query namespaces

The namespace list resource follows the same pattern:

list "kestra_namespace" "all_namespaces" {
  provider         = kestra
  limit            = 1000
  include_resource = true
}

Only namespaces that exist in Kestra are listed. A namespace that is only implied, because it holds flows or prefixes another namespace without having been created itself, is left out.

Generate configuration

Run the Terraform query with the variables used for the source Kestra instance:

terraform query \
  -var-file=tfvars \
  -generate-config-out=generated_kestra_imports.tf

Review the generated configuration before applying it to the target Terraform state.

Limiting the discovery

Use limit to cap the number of list results returned by Terraform:

list "kestra_flow" "sample" {
  provider = kestra
  limit    = 100
}

The limit value is a maximum result count for the query. It does not change the underlying Kestra resource data.

Resource contents

With include_resource = true, the provider performs the additional read required to populate the managed resource representation:

• kestra_flow retrieves the full flow including source content.
• kestra_namespace retrieves the full namespace representation.

Without include_resource, the query only needs the identity information and avoids those additional resource reads.