Hands-on 4: Variables and outputs

This exercise builds on Hands-on 3 in the same cumulative Terraform project.

Goal

Make the Snowflake configuration easier to reuse by introducing typed variables and useful outputs:

• user_suffix keeps the participant database unique.
• warehouse_size controls compute size.
• warehouse_auto_suspend controls automatic suspension.
• table_comment demonstrates configurable metadata.
• Outputs expose the current role, database, schema, fully qualified table, and warehouse.

Step 1: Add variables

Move the existing user_suffix variable to variables.tf if it is not there already, then add:

variable "warehouse_size" {
  description = "Snowflake warehouse size for the workshop."
  type        = string
  default     = "XSMALL"

  validation {
    condition     = contains(["XSMALL", "SMALL", "MEDIUM"], upper(var.warehouse_size))
    error_message = "warehouse_size must be XSMALL, SMALL, or MEDIUM."
  }
}

variable "warehouse_auto_suspend" {
  description = "Seconds of inactivity before the warehouse suspends."
  type        = number
  default     = 60
}

variable "table_comment" {
  description = "Comment stored on the workshop table."
  type        = string
  default     = "Table created during the Terraform workshop"
}

The existing terraform.tfvars continues to provide user_suffix. Optional values can be overridden there:

user_suffix            = "<your-username>"
warehouse_size         = "XSMALL"
warehouse_auto_suspend = 60

Step 2: Connect variables and outputs

Use the variables in the warehouse and table resources:

resource "snowflake_warehouse" "workshop" {
  name                = "WORKSHOP_WH_${upper(var.user_suffix)}"
  warehouse_size      = var.warehouse_size
  auto_suspend        = var.warehouse_auto_suspend
  auto_resume         = true
  initially_suspended = true
}

Add an output for the fully qualified table name:

output "qualified_table_name" {
  description = "The fully qualified name of the managed table."
  value       = "${snowflake_database.workshop.name}.${snowflake_schema.lab.name}.${snowflake_table.orders.name}"
}

The resource references are implicit dependencies. The table depends on the database and schema because it references their names; no depends_on is required.

Step 3: Plan with defaults and an override

terraform fmt
terraform validate
terraform plan
terraform output

Try a plan with a variable override without editing files:

terraform plan -var='warehouse_size=SMALL'

Review how Terraform proposes the warehouse change. Do not apply the override unless the workshop exercise explicitly requires it.

Common pitfalls

Terraform rejects a variable value

Read the validation error. Use an allowed warehouse size and a valid user suffix.

An output is unknown during plan

Some Snowflake values are known only after the provider reads or creates the object. This is expected; inspect the value with terraform output after apply.

depends_on is added unnecessarily

Prefer direct resource references when a dependency is expressed by a value such as a database or schema name.

Checkpoint

• Observe the precedence of defaults, terraform.tfvars, and -var command-line overrides.
• Identify each implicit dependency from the HCL references.
• Keep warehouse defaults conservative to avoid unnecessary Snowflake compute cost.