Densify

This module is the interface between Densify machine learning analytics and Terraform templates. It enables two operations:

• auto-tagging of cloud instances and containers based on Densify’s optimization analysis (making them “self-aware”)
• automated optimization of instance families/sizes, and container resource requests/limits (making them “self-optimizing”)

This integration is based on densify.auto.tfvars, which is automatically generated by Densify and contains operational intelligence, analysis findings, and optimization recommendations for each cloud instance or container in scope. This module unpacks data from this structure, making it available to Terraform as variables, enabling precise specification of resources to optimally match the learned patterns of behavior. The result is next-generation resource optimization with the elimination of hard-coded resource specifications.

• Requirements
• Usage
• Examples
• Inputs
• Outputs
• License

Requirements

• Densify account, which is provided with a Densify subscription or free trial (www.densify.com/service/signup)

Usage

module"optimization-as-code" {
source="densify-dev/optimization-as-code/null"densify_recommendations="${var.densify_recommendations}"densify_fallback="${var.densify_fallback}"densify_unique_id="${var.name}"
}

Examples

• AWS ASG
• AWS EC2
• AWS RDS
• Azure VM
• GCP Instance
• Kubernetes Pod
• Kubernetes Replication Controller
• Kubernetes Deployment

Inputs

Name	Description	Type	Default	Required
densify_recommendations	Map variable containing recommendations from Densify	Map	-	Yes
densify_fallback	The map default values used if Densify recommendations are not available	Map	-	Yes
densify_unique_id	The unique identifier of the system to be updated	String	-	Yes

Outputs

Name	Description
current_type	Current instance size and family.
recommended_type	Densify recommended instance size and family.
savings_estimate	The potential monthly savings from modifying the current instance to the Densify recommended instance.
predicted_uptime	The predicted percentage of CPU utilization hours over the duration of a month.
implementation_method	The system configured to allow automated or manual recommendations.
current_cpu_request	The current CPU request configured for the Kubernetes container.
recommended_cpu_request	The recommended CPU request for the Kubernetes container.
current_cpu_limit	The current CPU limit configured for the Kubernetes container.
recommended_cpu_limit	The recommended CPU limit for the Kubernetes container.
current_mem_request	The current memory request configured for the Kubernetes container.
recommended_mem_request	The recommended memory request for the Kubernetes container.
current_mem_limit	The current memory limit configured for the Kubernetes container.
recommended_mem_limit	The recommended memory limit for the Kubernetes container.
instance_type	The instance size and family to be implemented, which is either the current size or the Densify recommendation, depending on the automation policy and the approval status (if approval is enabled).
cpu_request	The CPU request to be implemented, which is either the current CPU request or the Densify recommendation, depending on the automation policy and the approval status (if approval is enabled).
cpu_limit	The CPU limit to be implemented, which is either the current CPU limit or the Densify recommendation, depending on the automation policy and the approval status (if approval is enabled).
mem_request	The memory request to be implemented, which is either the current memory request or the Densify recommendation, depending on the automation policy and the approval status (if approval is enabled).
mem_limit	The memory limit to be implemented, which is either the current memory limit or the Densify recommendation, depending on the automation policy and the approval status (if approval is enabled).
min_group_current	The current minimum group size of the ASG.
min_group_recommended	The recommended minimum group size for the ASG.
max_group_current	The current maximum group size of the ASG.
max_group_recommended	The recommended maximum group size for the ASG.
min_group	The minimum group size to be implemented, which is either the current minimum group size or the Densify recommended minimum group size depending on the automation policy and approval status (if approval is enabled).
max_group	The maximum group size to be implemented, which is either the current maximum group size or the Densify recommended maximum group size depending on the automation policy and approval status (if approval is enabled).
avg_inst_count_current	The current average number of instances running in the ASG.
avg_inst_count_recommended	The predicted average number of instances running in the ASG if the recommendations are implemented.
current_desired_capacity	The current desired number of instances running in the ASG.
desired_capacity	The desired capacity to be implemented, which is either the current desired capacity or the recommended average instance count (rounded down) depending on the automation policy and approval status (if approval is enabled).
display_name	The display name of the container, which is a combination of the pod\service name and the container name.
cluster	The name of the cluster with the running container.
namespace	The namespace for the running container.
pod_service	The pod or service name that the container is running in. If the container is running in a single pod, then the pod name is displayed. If the container is part of a service, such as a Replication Controller or Replica Set, then the service name is displayed.
container	The name of the container.
controller_type	The type of the controller.
current_size	The current number of copies of the controller that are running. The value will be 1 for individual pods.
approval_type	The type of approval for this system (NA, All, Specific size).
recommendation_type	The type of recommendation (Upsize, Downsize, Resize, Just Right).
effort	The effort rating for implementing the recommendation.
policy	The policy is used by the analysis.
power_state	The power state of the instance.
defer_recommendation	Defer recommending an instance type change until reserved instance coverage for this instance expires.
defer_until	If deferRecommendation = yes, then the expiry date of the RI providing coverage for the current instance is returned.

License

Apache 2 Licensed. See LICENSE for full details.