For the latest version of this help, see: https://help.apptio.com/en-us/cloudability/product/k8s-container-rightsizing.htm
Apptio Cloudability Container Rightsizing generates recommendations to help organizations optimize their Kubernetes container deployments. This article outlines capabilities of the Container Rightsizing functionality as part of an overall cloud financial optimization strategy.
Accessing Container Rightsizing
Container Rightsizing leverages the same reporting agent and data infrastructure used for the Container Cost Allocation functionality. If your organization is already leveraging the Cloudability Container Cost Allocation functionality, you will not need to do anything to utilize rightsizing.
If this is your organization's first time leveraging Apptio Cloudability Kubernetes functionality, you will need to provision your Kubernetes cluster(s) by deploying the opensource agent. For instructions on provisioning your environment, please visit the knowledge base article here. If you experience any issues deploying agents or have any questions, please contact support for assistance.
Accessing Container Rightsizing
To access the Container Rightsizing, select Rightsizing from within the Optimize menu on the left side primary navigation.
Like other areas of rightsizing the Container Rightsizing tab is organized with options for generating recommendations at the top of the page, followed by key metrics summaries and then the main table containing individual recommendations.
At the top of all rightsizing service tabs you will find recommendation preferences including:
- On-Demand - This basis generates Savings values based purely on On-Demand Pricing. While custom pricing is included in on-demand cost basis, any potential impact from commitments (Reserved Instances (RIs) or Savings Plans (SPs)) is not included.
- Effective - This basis takes into account the historical impact of commitments and in a similar manner to the Cost(Amortized) metric in that all associated upfront and recurring costs are included. In other words, this was the Effective cost of running your current configuration. For the recommended “New” instance type, the cost figures are based on the On-Demand prices. This is because the New configuration may not benefit from RIs or SPs. This comparison is the more conservative option. Even if you inadvertently move away from RIs/SPs, your new overall rate will still be better. As a result, the overall savings reported using this methodology will sometimes be lower. Custom pricing will be applied to these figures if applicable.
The account selection option enables users to filter recommendations down to view those related only to a specific account.
The timeline selection enables organizations to choose between using the previous 10 or 30 days of resources usage as the basis for generating recommendations.
Note: Visit the blog for tips on when to use 10 vs. 30 days of usage to generate recommendations.
Account Based Views
As with the Container Cost Allocation functionality, Container Rightsizing capability users can leverage account based views to filter recommendations down to a subset of their organizations accounts to view at any time.
Applying Filters With Links
- Click one or more of the blue hyperlinked values from any column in the main table.
- Your filters will then be added to the Filter Configurator above.
You must use the Filter Configurator to remove filter parameters.
NOTE: Only one value/parameter from each column may be selected at a time.
Using the Filter Configurator
To access the Filter Configurator, click the Pencil/Edit icon that appears when hovering above Filters in the Controls Sub-Navigation.
To add a filter, click Add Filter, and apply:
- a parameter (Measure)
- a functional/logical test (Operator)
- a Value to test against
When you are finished, click Submit.
To remove a filter, open the Filter Configurator and click the X to the right of the filter you want to remove.
As with on tabs for other services, summary KPIs for Container Rightsizing recommendations are detailed in four panes above the main table. KPIs are respective of views and filters so that when filters and views are applied, the summary KPIs will include only those containers that conform.
Summary KPIs include:
- Total Spend - the total current allocated spend across all containers instrumented in the Cloudabiltiy platform.
- Estimated Idle Savings - the estimated total savings for all ‘terminate’ recommendations displayed.
- Estimated Rightsizing Savings - the estimate total potential savings achievable for all 'rightsize' recommendations.
- Estimated Optimized Spend - the estimated total spend across all displayed containers after recommendations are applied.
Note: Rightsizing containers will adjust the cost allocation for the containers and pods with that container specification but ordinarily will not affect your bill. Rightsizing the container specification will make capacity available in your deployment, with which organizations can either use that space for other workloads or rightsize the underlying compute instance. Rightsizing the underlying compute instance will have a positive impact to an organization's bill.
Resources are sorted by cost savings by default. Start at the top of the list to see the underutilized resources with the highest potential savings.
Viewing Recommendation Details
For each line represented in the main table, there are up to two rightsizing recommendations generated. The main table displays cost, expected savings and resourse values for the suggested recommendation.
Accessing the details pane will allow you to view specific information about individual recommendations. Clicking the Details button, or the three dots at the right of any row in the main table will open the Details pane.
Within the Details pane, at least one recommendation is presented with the lowest risk option being the default selection.
Each option includes request and limit recommendations for Memory and CPU for a given container specification. Graphs for individual resources display the current resource allocation (in red) and consumption (in blue) along with the recommended settings (in yellow).
For additional information on managing your Kubernetes Container deployments with Apptio Cloudability please visit: