Provider instance management

Ensure proper configuration of provider instances when using environment-specific settings. When using for_each with providers, carefully manage the lifecycle of provider instances to avoid warnings or errors during resource operations.

Key considerations:

1. Provider instances created with for_each must remain available for the entire lifecycle of any resources using them, including during destruction
2. When resources also use for_each, ensure their instance keys don’t exactly match the provider’s to allow proper resource removal
3. Use selective filtering when creating resource instances to maintain needed provider instances

# Good practice: Filter resources but maintain provider instances
variable "aws_regions" {
  type = map(object({
    vpc_cidr_block = string
  }))
}

provider "aws" {
  alias    = "by_region"
  for_each = var.aws_regions

  region = each.key
}

resource "aws_vpc" "private" {
  # Filter var.aws_regions to include only non-null elements
  # allowing provider instances to remain while removing resources
  for_each = {
    for region, config in var.aws_regions : region => config
    if config != null
  }
  provider = aws.by_region[each.key]

  cidr_block = each.value.vpc_cidr_block
}

This pattern allows you to “disable” resources in specific regions by setting their config to null while maintaining the provider instance needed for cleanup operations. Without this approach, OpenTofu would generate warnings when trying to destroy resources whose provider instances have been removed.