Accessing other resources created by the same for_each set - terraform

I'm trying to create an az ad app and credential for each entry in a locals set.
The objects in the locals set have values that are needed for both resources, but my issue is the credentials resource needs values from both the locals object as well as the ad application.
This would be easy normally, but I am using a for_each which is complicated, and the value of each for the credential resource is the ad application. Is there any way I can get access to the each of az app resource but from the credential resource?
locals {
github_repos_with_apps = {
tftesting_testing = {
repo = "tftesting-testing"
environment = "tfplan"
}
}
}
resource "azuread_application" "aadapp" {
for_each = local.github_repos_with_apps
display_name = join("-", ["github-actions", each.value.repo, each.value.environment])
owners = [data.azuread_client_config.current.object_id]
}
resource "azuread_application_federated_identity_credential" "cred" {
for_each = azuread_application.aadapp
application_object_id = each.value.object_id
display_name = "my-repo-deploy"
description = "Deployments for my-repo"
audiences = ["api://AzureADTokenExchange"]
issuer = "https://token.actions.githubusercontent.com"
subject = "repo:my-org/${each.value.<something?>.repo}:environment:${each.value.<something?>.environment}"
}
In the snippet above I need the cred resource to access aadapp.object_id but also reference the locals value in order to get rep and environment. Since both cred and aadapp both use for_each the meaning of each.value changes. I'd like to reference the each.value of aadapp from cred.
My problem line is the subject value in the cred resource:
subject = "repo:my-org/${each.value.<something?>.repo}:environment:${each.value.<something?>.environment}"
I think I may have to use modules to accomplish this, but I feel there is a quicker way, like being able to store a temporary value on aadapp that would let me reference it.

After scouring some examples I did find out how to achieve this.
If I change all resources to use for_each = local.github_repos_with_apps, I can then use 'each.key` as a lookup to get the other associated resources like so:
application_object_id = resource.azuread_application.aadapp[each.key].object_id
This allows the cred resource to reference the locals values directly
subject = "repo:my-org/${each.value.repo}:environment:${each.value.environment}"
Full code:
locals {
github_repos_with_apps = {
first_test : {
repo = "tftesting-testing"
environment = "tfplan"
}
second_test : {
repo = "bleep-testing"
environment = "tfplan"
}
}
}
resource "azuread_application" "aadapp" {
for_each = local.github_repos_with_apps
display_name = join("-", ["github-actions", each.value.repo, each.value.environment])
owners = [data.azuread_client_config.current.object_id]
lifecycle {
ignore_changes = [
required_resource_access
]
}
}
resource "azuread_application_federated_identity_credential" "cred" {
for_each = local.github_repos_with_apps
application_object_id = resource.azuread_application.aadapp[each.key].object_id
display_name = each.value.repo
description = "Deployments for my-repo"
audiences = ["api://AzureADTokenExchange"]
issuer = "https://token.actions.githubusercontent.com"
subject = "repo:my-org/${each.value.repo}:environment:${each.value.environment}"
}

Related

Terraform flatten and looping issue inside modules

I'm trying to create a module in Terraform for creating Azure resources and facing some issues. This module creates a resource group, subnet, vnet and Role bindings. I see that the below code creates the resources twice because of the loop. Does the for_each loop work in such a way that the entire resource or module block will be executed each time it loops? I'm new to Terraform and come from a Java background.
Also, ideally would like to use the flatten inside the module without locals possibly, any way to do that? Code is below.
locals {
groupsbyrole = flatten([
for roleName, groupList in var.testproject1_role_assignments : [
for groupName in groupList : {
role_name = roleName
group_name = groupName
}
]
])
}
module "testproject1" {
source = "C:\\Users\\ebo1h8h\\Documents\\Project\\Automation\\Terraform\\Code\\Azure\\modules\\sandbox-module"
short_name = "testproj"
# Resource Group Variables
az_rg_location = "eastus"
az_tags = {
Environment = "Sandbox"
CostCenter = "Department"
ResourceOwner = "Vikram"
Project = "testproj"
Role = "Resource Group"
}
address_space = ["10.0.0.0/16"]
subnet_prefixes = ["10.0.1.0/24"]
subnet_names = ["a-npe-snet01-sbox"]
vnet_location = var.az_rg_location
for_each = {
for group in local.groupsbyrole : "${group.role_name}.${group.group_name}}" => group
}
principal_id = each.value.group_name
role_definition_name = each.value.role_name
}
And here is the role_assignments variable
variable "testproject1_role_assignments" {
type = map(list(string))
default = {
"Contributor" = ["prod-azure-contrib-sbox", "gcp-org-network-engineering"],
"Owner" = ["gcp-org-cloud-delivery"]
}
}
The above code creates 12 resources when it should be only 6. The only was I was able to get around this is have the resource "azurerm_role_assignment" "role_assignment" as a separate module. Ideally, I want to pass the role assignments variable in each of the module to be created so that it creates a set of resources.
Any pointers on how to achieve that?
Thanks,
The docs state
If a resource or module block includes a for_each argument whose value is a map or a set of strings, Terraform will create one instance for each member of that map or set.
So in your scenario you are creating 3 instances of the module, whereas it sounds like you want to pass in the local.groupsbyrole object as a variable in the module and only attach the for_each to the resources you want multiple instances of.
Sidenote: You could simplify your local by adding group like below:
locals {
groupsbyrole = flatten([
for roleName, groupList in var.testproject1_role_assignments : [
for groupName in groupList : {
role_name = roleName
group_name = groupName
group = "${roleName}.${groupName}"
}
]
])
}
Tip: I find adding an output to see the shape of the object whilst developing can also be useful
output "test_output" {
value = local.groupsbyrole
}
Then when you run plan you will see your object
test_output = [
+ {
+ group = "Contributor.prod-azure-contrib-sbox"
+ group_name = "prod-azure-contrib-sbox"
+ role_name = "Contributor"
},
+ {
+ group = "Contributor.gcp-org-network-engineering"
+ group_name = "gcp-org-network-engineering"
+ role_name = "Contributor"
},
+ {
+ group = "Owner.gcp-org-cloud-delivery"
+ group_name = "gcp-org-cloud-delivery"
+ role_name = "Owner"
},
]

terraform - Iterate over two linked resources

I’m trying to write some code which would take an input structure like this:
projects = {
"project1" = {
namespaces = ["mynamespace1"]
},
"project2" = {
namespaces = ["mynamespace2", "mynamespace3"]
}
}
and provision multiple resources with for_each which would result in this:
resource "rancher2_project" "project1" {
provider = rancher2.admin
cluster_id = module.k8s_cluster.cluster_id
wait_for_cluster = true
}
resource "rancher2_project" "project2" {
provider = rancher2.admin
cluster_id = module.k8s_cluster.cluster_id
wait_for_cluster = true
}
resource "rancher2_namespace" "mynamespace1" {
provider = rancher2.admin
project_id = rancher2_project.project1.id
depends_on = [rancher2_project.project1]
}
resource "rancher2_namespace" "mynamespace2" {
provider = rancher2.admin
project_id = rancher2_project.project2.id
depends_on = [rancher2_project.project2]
}
resource "rancher2_namespace" "mynamespace3" {
provider = rancher2.admin
project_id = rancher2_project.project2.id
depends_on = [rancher2_project.project2]
}
namespaces are dependent on Projects and the generate id needs to be passed into namespace.
Is there any good way of doing this dynamically ? We might have a lot of Projects/namespaces.
Thanks for any help and advise.
The typical answer for systematically generating multiple instances of a resource based on a data structure is resource for_each. The main requirement for resource for_each is to have a map which contains one element per resource instance you want to create.
In your case it seems like you need one rancher2_project per project and then one rancher2_namespace for each pair of project and namespaces. Your current data structure is therefore already sufficient for the rancher2_project resource:
resource "rancher2_project" "example" {
for_each = var.projects
provider = rancher2.admin
cluster_id = module.k8s_cluster.cluster_id
wait_for_cluster = true
}
The above will declare two resource instances with the following addresses:
rancher2_project.example["project1"]
rancher2_project.example["project2"]
You don't currently have a map that has one element per namespace, so it will take some more work to derive a suitable value from your input data structure. A common pattern for this situation is flattening nested structures for for_each using the flatten function:
locals {
project_namespaces = flatten([
for pk, proj in var.projects : [
for nsk in proj.namespaces : {
project_key = pk
namespace_key = ns
project_id = rancher2_project.example[pk].id
}
]
])
}
resource "rancher2_namespace" "example" {
for_each = {
for obj in local.project_namespaces :
"${obj.project_key}.${obj.namespace_key}" => obj
}
provider = rancher2.admin
project_id = each.value.project_id
}
This produces a list of objects representing all of the project and namespace pairs, and then the for_each argument transforms it into a map using compound keys that include both the project and namespace keys to ensure that they will all be unique. The resulting instances will therefore have the following addresses:
rancher2_namespace.example["project1.mynamespace1"]
rancher2_namespace.example["project2.mynamespace2"]
rancher2_namespace.example["project2.mynamespace3"]
This seems to work too:
resource "rancher2_namespace" "example" {
count = length(local.project_namespaces)
provider = rancher2.admin
name = local.project_namespaces[count.index].namespace_name
project_id = local.project_namespaces[count.index].project_id
}

Azure Function AppSettings using Terraform and mutliple sources for map

so in summary I am specifically looking to maintain the app settings for my azure functions using two different sources,
the first source is a map of custom settings that will be maintained manually or through code which might have little change
The second source of app settings map are key secret uri's as per the code before, this enables the azure function to use secret references as configuration value.
I am trying to automate the process of retrieving a subset of secrets dynamically from keyvault and merging it into the custom map app settings that I define in code.
Question:
My ideal world would be that i update the list secretKeys and the map appSettingsSecretsMap get's dynamically created and then consumed by resource creation resource "azurerm_function_app" "functionApp_workerFunctions" in its appsettings. Does anyone have a idea of how I might achieve this a bit more dynamically?
My full code is as per below:
variable "secretKeys" {
type = list(string)
default = [
"TestDbPassword",
"TestDbUserId"]
}
data "azurerm_key_vault" "keyvault" {
name = "source-keyvault"
resource_group_name = "source-keyvault-rg"
}
data "azurerm_key_vault_secret" "kvSecrets" {
for_each = toset(var.secretKeys)
name = each.key
key_vault_id = data.azurerm_key_vault.keyvault.id
}
# Testing Access to secret
output "TestDbPassword" {
value = data.azurerm_key_vault_secret.kvSecrets["TestDbPassword"].id
}
#https://learn.microsoft.com/en-us/azure/app-service/app-service-key-vault-references
variabe "appSettingsSecretsMap" {
type = map
default = {
DBPassword = "#Microsoft.KeyVault(SecretUri=${data.azurerm_key_vault_secret.kvSecrets["TestDbPassword"].id})"
DBUserId = "#Microsoft.KeyVault(SecretUri=${data.azurerm_key_vault_secret.kvSecrets["TestDbUserId"].id})"
}
}
# Reference for appSettings https://learn.microsoft.com/en-us/azure/azure-functions/functions-app-settings
variable "appSettingsCustomMap" {
type = map
default = {
WEBSITE_RUN_FROM_PACKAGE = ""
FUNCTIONS_WORKER_RUNTIME = ""
APPINSIGHTS_INSTRUMENTATIONKEY = ""
#FUNCTIONS_EXTENSION_VERSION = "~1"
}
}
resource "azurerm_function_app" "functionApp_workerFunctions" {
name = "worker-function-${var.ENVIRONMENT}"
location = "XYZ-Example"
resource_group_name = "XYZ-Example"
app_service_plan_id = "XYZ-Example"
storage_account_name = "XYZ-Example"
storage_account_access_key = "XYZ-Example"
app_settings = merge(var.appSettingsMap, var.appSettingsSecretsMap)
}
For the custom settings, I think it's better to set it in variable manually. It has a lot of things with a little change. If you make it automated, I think it will be a little redundancy. Just do it as you show in the question.
For key fault, I recommend you use the locals block:
locals {
appSettingsSecretsMap = {
DBPassword = "#Microsoft.KeyVault(SecretUri=${data.azurerm_key_vault_secret.kvSecrets[0].id})"
DBUserId = "#Microsoft.KeyVault(SecretUri=${data.azurerm_key_vault_secret.kvSecrets[1].id})"
}
}
The data for the key vault secrets will return a list with the element like this:
So you cannot quote it with the secret name. Just do it as what I show you.

Terraform for each loop or for loop not working

I am trying to set role for azure container registry for multiple service principals
variable "custom_role_list" {
type = list(object ({ service_principal_id = string, role = string }) )
}
When i try to set it from resource module, which I am not sure is the correct way?
resource "azurerm_role_assignment" "ad_sp_role_assignment" {
scope = azurerm_container_registry.acr.id
for_each = var.custom_role_list
role_definition_name = each.value.role
principal_id = each.value.service_principal_id
}
Essentially I am trying to set the azure container registry to work with multiple service principal with specific access roles.
Following is the var definition.
custom_role_list = [
{
service_principal_id = aserviceprincipal.id
role = "Contributor"
},
{
service_principal_id = bserviceprincipal.id
role = "Contributor"
}
]
When I execute it I get the following error.
Error: Invalid for_each argument
on ../modules/az-acr/main.tf line 46, in resource "azurerm_role_assignment" "ad_sp_role_assignment":
46: for_each = var.custom_role_list
The given "for_each" argument value is unsuitable: the "for_each" argument
must be a map, or set of strings, and you have provided a value of type list
of object.
Please if someone can guide will be very much helpful. thanks!
As the error suggests, for_each only supports maps and sets when used with a resource. You're trying to use a list of objects.
Instead, perhaps your variable can be simply of type map, where each service principle is a key and its corresponding role is the value. For example:
variable "custom_role_list" {
type = map
}
The variable definition:
custom_role_map = {
aserviceprincipal.id = "Contributor"
bserviceprincipal.id = "Contributor"
}
And finally use for_each:
resource "azurerm_role_assignment" "ad_sp_role_assignment" {
for_each = var.custom_role_map
scope = azurerm_container_registry.acr.id
role_definition_name = each.value
principal_id = each.key
}
You might find this blog post to help you with using loops and conditionals with Terraform.
You can use a for_each loop with your list of objects by adapting your code to the following:
variable "custom_role_list" {
type = list(object({
service_principal_id = string
role = string
}))
default = [
{
service_principal_id= "27d653c-aB53-4ce1-920",
role = "Contributor"
},
{
service_principal_id= "57d634c-aB53-4ce1-397",
role = "Contributor"
}
]
}
resource "azurerm_role_assignment" "ad_sp_role_assignment" {
for_each = {for sp in var.custom_role_list: sp.service_principal_id => sp}
scope = azurerm_container_registry.acr.id
role_definition_name = each.value.service_principal_id
principal_id = each.value.role
}

How do I make terraform skip that block while creating multiple resources in loop from a CSV file?

Hi I am trying to create a Terraform script which will take inputs from the user in the form of a CSV file and create multiple Azure resources.
For example if the user wants to create: ResourceGroup>Vnet>Subnet in bulk, he will provide input in CSV format as below:
resourcegroup,RG_location,RG_tag,domainname,DNS_Zone_tag,virtualnetwork,VNET_location,addressspace
csvrg1,eastus2,Terraform RG,test.sd,Terraform RG,csvvnet1,eastus2,10.0.0.0/16,Terraform VNET,subnet1,10.0.0.0/24
csvrg2,westus,Terraform RG2,test2.sd,Terraform RG2,csvvnet2,westus,172.0.0.0/8,Terraform VNET2,subnet1,171.0.0.0/24
I have written the following working main.tf file:
# Configure the Microsoft Azure Provider
provider "azurerm" {
version = "=1.43.0"
subscription_id = var.subscription
tenant_id = var.tenant
client_id = var.client
client_secret = var.secret
}
#Decoding the csv file
locals {
vmcsv = csvdecode(file("${path.module}/computelanding.csv"))
}
# Create a resource group if it doesn’t exist
resource "azurerm_resource_group" "myterraformgroup" {
count = length(local.vmcsv)
name = local.vmcsv[count.index].resourcegroup
location = local.vmcsv[count.index].RG_location
tags = {
environment = local.vmcsv[count.index].RG_tag
}
}
# Create a DNS Zone
resource "azurerm_dns_zone" "dnsp-private" {
count = 1
name = local.vmcsv[count.index].domainname
resource_group_name = local.vmcsv[count.index].resourcegroup
depends_on = [azurerm_resource_group.myterraformgroup]
tags = {
environment = local.vmcsv[count.index].DNS_Zone_tag
}
}
To be continued....
The issue I am facing here what is in the second resource group, the user don't want a resource type, suppose the user want to skip the DNS zone in the resource group csvrg2. How do I make terraform skip that block ?
Edit: What I am trying to achieve is "based on some condition in the CSV file, not to create azurerm_dns_zone resource for the resource group csvrg2"
I have provided an example of the CSV file, how it may look like below:
resourcegroup,RG_location,RG_tag,DNS_required,domainname,DNS_Zone_tag,virtualnetwork,VNET_location,addressspace
csvrg1,eastus2,Terraform RG,1,test.sd,Terraform RG,csvvnet1,eastus2,10.0.0.0/16,Terraform VNET,subnet1,10.0.0.0/24
csvrg2,westus,Terraform RG2,0,test2.sd,Terraform RG2,csvvnet2,westus,172.0.0.0/8,Terraform VNET2,subnet1,171.0.0.0/24
you had already the right thought in your mind using the depends_on function. Although, you're using a count inside, which causes from my understanding, that once the first resource[0] is created, Terraform sees the dependency as solved and goes ahead as well.
I found this post with a workaround which you might be able to try:
https://github.com/hashicorp/terraform/issues/15285#issuecomment-447971852
That basically tells us to create a null_resource like in that example:
variable "instance_count" {
default = 0
}
resource "null_resource" "a" {
count = var.instance_count
}
resource "null_resource" "b" {
depends_on = [null_resource.a]
}
In your example, it might look like this:
# Create a resource group if it doesn’t exist
resource "azurerm_resource_group" "myterraformgroup" {
count = length(local.vmcsv)
name = local.vmcsv[count.index].resourcegroup
location = local.vmcsv[count.index].RG_location
tags = {
environment = local.vmcsv[count.index].RG_tag
}
}
# Create a DNS Zone
resource "azurerm_dns_zone" "dnsp-private" {
count = 1
name = local.vmcsv[count.index].domainname
resource_group_name = local.vmcsv[count.index].resourcegroup
depends_on = null_resource.example
tags = {
environment = local.vmcsv[count.index].DNS_Zone_tag
}
}
resource "null_resource" "example" {
...
depends_on = [azurerm_resource_group.myterraformgroup[length(local.vmcsv)]]
}
or depending on your Terraform version (0.12+ which you're using guessing your syntax)
# Create a resource group if it doesn’t exist
resource "azurerm_resource_group" "myterraformgroup" {
count = length(local.vmcsv)
name = local.vmcsv[count.index].resourcegroup
location = local.vmcsv[count.index].RG_location
tags = {
environment = local.vmcsv[count.index].RG_tag
}
}
# Create a DNS Zone
resource "azurerm_dns_zone" "dnsp-private" {
count = 1
name = local.vmcsv[count.index].domainname
resource_group_name = local.vmcsv[count.index].resourcegroup
depends_on = [azurerm_resource_group.myterraformgroup[length(local.vmcsv)]]
tags = {
environment = local.vmcsv[count.index].DNS_Zone_tag
}
}
I hope that helps.
Greetings

Resources