Invalid for_each: log_arn_chunks uses unknown values at plan time (module fails with AWS Provider 6.x)

[ninos-atg]

I'm getting a plan-time failure in version v10.0.1 of the module caused by a for_each using unknown values:

Error: Invalid for_each argument
local.log_arn_chunks is a list of list of string, known only after apply

This happens here:

for_each = {
  for idx, chunk in local.log_arn_chunks : idx => chunk if length(chunk) > 0
}

Additionally, the module pins multiple incompatible AWS provider versions, causing:

Failed to query available provider packages: no available releases match the given constraints >=5.49.0 ... <7.0.0

Please update the module to support Terraform 1.6+ and AWS Provider 6.x.

[mlcooper]

The module currently supports terraform >= 1.0.0, and the aws provider, ">= 6.11.0, < 7.0.0". Please reference the terraform.tf file and the modules/kms_secrets/terraform.tf file. There are no other restrictions on terraform or the aws provider version.

[ninos-atg]

So is this module ("disney/kinesis-firehose-splunk/aws" v10.0.1 ) support aws version = "6.21.0"?

[ninos-atg]

Maybe I can explain it a little more clearly:

We’re using this module:

source = "disney/kinesis-firehose-splunk/aws"
version = v10.0.1

and we’re hitting this error during terraform plan in our pipeline:
`Error: Invalid for_each argument

on main.tf line 225, in resource "aws_iam_policy" "log_access_chunks":
225: for_each = { for idx, chunk in local.log_arn_chunks : idx => chunk if length(chunk) > 0 }

local.log_arn_chunks is a list of list of string, known only after apply

The "for_each" map includes keys derived from resource attributes that cannot
be determined until apply...`

What changed before the issue started

The failure started right after we enabled default_tags in our AWS providers.

Before:

`provider "aws" {
region = "eu-north-1"

default_tags {
tags = {
map-migrated = "migNXHGIDE7GL"
}
}
}`

after:
`provider "aws" {
region = "eu-north-1"

default_tags {
tags = {
map-migrated = "migNXHGIDE7GL"
Environment = var.environment
Application = var.application
awsApplication = "NA"
CostCenter = ""
WorkStream = ""
"sagemaker:user-profile-arn" = "NA"
}
}
}`

Root cause

Inside the module, for_each keys are derived from:

local.log_arn_chunks

In our case, these values become known only after apply because:

• Our log groups are provided dynamically (data.external)
• AWS provider default_tags adds implicit dependencies / changes evaluation order
• Result: local.log_arn_chunks is no longer fully known at plan time

Terraform requires for_each keys to be known during plan, so the module fails.