I want to get a map using terraform data which has the following structure:
{
instance-id(e.g. i-xxxx): ValueOfNameTag(e.g. MongoDB)
}
I can get the instance details individually like below
data "aws_instance" "test" {
instance_id = "i-0ecb3c2fb4436b6fe"
}
But don't know hot to do it for all and create a map out of it.
CodePudding user response:
This is essentially a three step process. The first step (already exists according to comments) is to retrieve the desired list of instance IDs:
data "aws_instances" "this" {
# arguments to select specific subset of instances
}
The list(string)
of instances IDs will be stored in the exported resource attribute ids
.
The second step is to retrieve the tags of each individual instance:
data "aws_instance" "this" {
# iterate over subset of desired instances by ID
for_each = toset(data.aws_instances.this.ids)
instance_id = each.value
}
The tags for each instance will be stored in the exported resource attribute tags
, and therefore can be accessed via the namespace data.aws_instance.this["<instance id>"].tags
as per normal in Terraform HCL2 DSL.
The third step is to perform a data transformation to construct a map
where the key is the instance ID, and the value of a tag is the value. The example given in the question was for the value of the Name
tag, so that is what we will use here in the example:
locals {
instance_name = { for id, attributes in data.aws_instance.this : id => attributes.tags["Name"] }
}
and the desired map
of instance IDs to their respective Name
tag will be stored in local.instance_name
as expected. You could alternatively iterate over the data.aws_instances.this.ids
list(string)
in a for
lambda expression and map to the exported attributes from the data.aws_instance.this[each.value].tags["Name"]
, but the above algorithm is slightly cleaner and slightly more efficient.