aws_cloudfront_response_headers_policy: added server_timing_headers_config attribute

PR is here

Shipped in Provider version 4.16.0.

Adding new server_timing_headers_config configuration block for aws_cloudfront_response_headers_policy resource. It will allow us to configure Server Timing headers within Cloudfront.

resource "aws_cloudfront_response_headers_policy" "test" {
  name = "my-policy"
	
  server_timing_headers_config {
    enabled       = true
    sampling_rate = 20.5 
  }
}

AWS API and userguide docs refs:

• API_ResponseHeadersPolicyServerTimingHeadersConfig
• API_CreateResponseHeadersPolicy
• Response Headers Policies
• Creating Response Headers Policies

Lessons learned

Update both resource and data_source

If you are going to introduce some new attributes for AWS resource it will be a good practice to also add the same attribute for data_source. This is useful as you will have the attribute populated for your data_source when you are trying to just read some already existing resources.

Please pay attention, that all the attributes you define for your data_source MUST have Computed: true defined for them:

func DataSourceResponseHeadersPolicy() *schema.Resource {
  return &schema.Resource{
  Read: dataSourceResponseHeadersPolicyRead,
  
  Schema: map[string]*schema.Schema{
      ...
      "server_timing_headers_config": {
          Type:     schema.TypeList,
          Computed: true,
          Elem: &schema.Resource{
              Schema: map[string]*schema.Schema{
                  "enabled": {
                      Type:     schema.TypeBool,
                      // here 
                      Computed: true,
                  },
                  "sampling_rate": {
                      Type:     schema.TypeFloat,
                      // and here
                      Computed: true,
                  },
              }},
          },
      },
  }
}

Always try to follow AWS API (SDK) objects structure as much as possible

Based on AWS API docs and AWS SDK, Server Timing headers configuration object contains two fields:

type ResponseHeadersPolicyServerTimingHeadersConfig struct {
    Enabled *bool `type:"boolean" required:"true"`
    SamplingRate *float64 `type:"double"`
}

From the first look it might appear to you that having Enabled attribute is redundant as it does not make any sense to set it to false - you can just remove Terraform configuration block for server timing headers config.

My first version of the code had an if statement for doing this enablement/disablement:

func expandResponseHeadersPolicyServerTimingHeadersConfig(tfMap map[string]interface{}) *cloudfront.ResponseHeadersPolicyServerTimingHeadersConfig {
    ...
    apiObject := &cloudfront.ResponseHeadersPolicyServerTimingHeadersConfig{}
    
    if v, ok := tfMap["sampling_rate"].(float64); ok && v != 0 {
        // only set some values if there is "sampling_rate" provided
        apiObject.Enabled = aws.Bool(v)
        apiObject.SamplingRate = aws.Float64(v)
    } else {
        // disable if block is missing in Terraform code
        apiObject.Enabled = aws.Bool(false)
    }

	return apiObject
}

However, after review Terraform Contribution guidelines I have figured out that it’s better to follow API object structure as much as possible.

That means, to have separate fields for enabled and sampling_rate:

func expandResponseHeadersPolicyServerTimingHeadersConfig(tfMap map[string]interface{}) *cloudfront.ResponseHeadersPolicyServerTimingHeadersConfig {
  ...
  apiObject := &cloudfront.ResponseHeadersPolicyServerTimingHeadersConfig{}
  
  if v, ok := tfMap["enabled"].(bool); ok {
      apiObject.Enabled = aws.Bool(v)
  }
  
  if v, ok := tfMap["sampling_rate"].(float64); ok && v != 0 {
      apiObject.SamplingRate = aws.Float64(v)
  }
  
  return apiObject
}