Optimizing multi-cloud costs with recommendations
BMC Helix Cloud Cost provides actionable intelligence to handle efficiency-related issues in your infrastructure through a set of recommendations. By implementing these recommendations, you can reduce your multi-cloud costs and optimally utilize the infrastructure.
A recommendation is a suggestion based on best practices about how to improve the cost efficiency of your infrastructure. Recommendations are available for the following cloud providers:
- Amazon Web Services (AWS)
- Microsoft Azure (Azure)
Recommendations are generated by analyzing the utilization metrics of an instance. To generate recommendations you must configure the connector to collect the cost and utilization data. For a list of connectors, see Collecting-data.
The following video (1:39) provides information about the recommendations available in BMC Helix Cloud Cost.
Each recommendation comprises the following elements:
- Brief overview of the issue
- One or more suggested actions to resolve the issue
- Estimated monthly savings that you can achieve by performing the suggested actions
The recommendations are available in two categories. The following table lists the categories and the goals that the recommendations help you achieve:
Category | Goals |
|---|---|
Resource usage |
|
Reserved instance1 | Determine the number of AWS reserved instances that can be purchased to optimize your current costs |
1 - Available only if you are using the AWS cloud provider.
Resource usage recommendations
The recommendations under the Resource usage category are further classified as follows:
Idle VM
Idle VMs are virtual machines that remain running although they are not used. For example, VMs were provisioned for testing a product feature but after the testing task completed, the VMs were not decommissioned. The public cloud providers continue to charge you for such idle VMs.
The Idle VM recommendation identifies virtual machines in your multi-cloud infrastructure that are idle or unused for a specific number of days, and also suggests actions to terminate them.
For information about how the recommendation is generated, see Analyzing-the-idle-VMs-in-your-multi-cloud-infrastructure.
Overallocated VM
Overallocated VMs are virtual machines with underutilized resources like CPU and memory. Users often end up provisioning VMs with excess capacity because they are not sure of the correct size that can meet their requirements. The public cloud providers charge you for the total capacity whether you use it or not.
The Overallocated VM recommendation identifies virtual machines in your multi-cloud infrastructure with underutilized resources, and also suggests the ideal size of the VMs; so that you can reduce your costs. A VM is considered to be overallocated if the spare capacity of any of its resources is less than the specified threshold value when projected for 30 days in the future. The suggested actions in the recommendation provide the estimated savings that you can achieve by resizing the VMs.
For information about how the recommendation is generated, see Analyzing-the-overallocated-VMs-in-your-multi-cloud-infrastructure.
Idle VM (Power-Off)
Idle VMs are virtual machines that display a regular usage pattern with a significant idle time. The Power-Off VMs recommendation identifies virtual machines in your multi-cloud infrastructure that are idle or unused for a specific number of days, and also suggests actions to power-off these VMs for specific days of the month and duration of time to achieve cost savings.
For information about how the recommendation is generated, see Analyzing-power-on-off-recommendations-for-VMs.
Reserved instance recommendation
AWS reserved instances offer a significant discount compared to the on-demand pricing. The reserved instance recommendation specifies the number of reserved instances that you must purchase to achieve cost savings. These recommendations are calculated based on the past usage of your existing instances, which are charged at on-demand rates, and checks which of these on-demand instances can be reserved.
For information about how the recommendation is generated, see Analyzing-recommendations-for-purchasing-AWS-reserved-instances.