Get recommendations for buying GCP Resource-based Commitments

Edit online

The Cloudability GCP CUD recommendation interface will allow you to understand your current usage patterns, include your future usage and business strategies, and balance your risk profile to maximize your discounts by purchasing the best CUDs for your future requirements.

CUD fundamentals

Committed use discounts (CUDs) provide discounted prices in exchange for your commitment to use a minimum level of resources for a specified term. The discounts are flexible, cover a wide range of resources, and are ideal for workloads with predictable resource needs.

When you purchase Google Cloud committed use discounts, you commit to a consistent amount of usage for a one- or three-year period. You receive a discounted rate on the usage your commitment covers, but any usage over the committed amount is charged at the on-demand rate.

How to get the Most Accurate Recommendations:

To use Cloudability to provide the most accurate recommendations for your future usage, take the following approach:

Setup your organization: Prepare your usage in Cloudability so that the right dimensions are available in the GCP CUD recommendation engine. This will enable you to filter and adjust your usage on these dimensions for analysis.

Select the data for analysis: Chose the best historical time period that will match your future usage patterns from within the current and previous months usage.

Refine the data: Using the dimensions available, filter and remove any data that you would not want to cover with a commitment, for example removing temporary usage such as test workloads or workloads that are being decommissioned.

Generating recommendations

Setup your organization

The first step of using the CUD recommendation component is to setup your organization. This ensures that the data in the recommendation engine has the required dimensions to be able to filter on.

As these settings apply to data across your organization additional privileges are required to access the preferences page. Add either CommitmentPreferencesFeatureViewOnly or CommitmentPreferencesFeatureFullAccess for access.

  • ViewOnly will allow a user to view what preferences are currently configured and will be applied in the next processing run.
  • FullAccess will allow a user to modify the preferences to be applied at the next processing run.

To setup your organization, go to Settings and select Commitment Preferences . Here you will be able to configure up to 10 Cloudability dimensions to apply to your usage. Select dimensions from your configured Tags & Labels, Business Mappings, or Account Groups.

For best performance, choose dimensions that provide a moderate level of variability – such as business units or application names. Dimensions that have a large amount of variability such as resource name will lead to significant performance degradation. If a configured dimension is not available – it is likely due to excessive variability.

Any changes to your preferences will be implemented in the next processing cycle, allow up to 24 hours for any changes to take effect, and be available in the CUD recommendations page.

Dimensions are regularly checked for variability and will be added/removed from the list of available dimensions. If a dimension becomes too variable it will be removed and no longer available in your data set.

Select the Data for Analysis

The second step is to select the data for analysis, ensure you select previous usage that reflects your future usage as closely as possible. Under the Optimize menu, select Reserved Instance Planner , then select the GCP tab:

Select the required Duration of the commitment, then using the date picker select a time range from within the current and previous month for analysis, then select one or more accounts to analyze:

Note:

Depending on your time zone and when the vendors last update was, the current and/or previous days data may be incomplete and impact results. Ensure you have data throughout the entire selected dates for the most accurate results.

Refine the Data

The third step is to refine the data and remove any usage that should not be included in the commitment analysis. Usage such as test workloads or workloads that will be decommissioned can be removed.

Add one or more filters based on your pre-configured dimensions:

Filers are applied either pre or post calculation. Pre filters provide more control and granularity on the data going into the analysis. Post-filters remove results from being displayed after the analysis.

  • Pre-filters: All dimensions defined in the Commitment preferences page, machine family, machine type.
  • Post-filters: Commitment type, region, cpu quantity, memory quantity, total savings, percent savings.

View the recommendations

You can now view the default recommendations which provide the lowest overall cost with a balanced risk. The main KPI’s are across the top, and the individual recommendations are in the table below.

The table is sorted to ensure the highest value & lowest risk CUDs are presented at the top. It is sorted by % savings descending, then by $ savings descending. You can change this order by selecting a columns heading to change the sort priority by that heading, either ascending or descending.

  • Total Savings (Estimated Net Savings): The estimated amount of money saved if the recommendations were purchased.
  • Total Savings Percentage (Estimated Savings Rate): This is the total savings amount, divided by the original cost of usage without CUDs applied. This original cost would include any previously applied SUDs, and would exclude any usage that was filtered by the user.

Tailoring Recommendations

The final step in the process is to tailor the recommendations to your future business requirements and balance the risk. In this step you observe your usage trends and analyze different levels of commitment and the results. Select the most appropriate level of commitment coverage for your future business requirements.

Select the details icon for the specific recommendation:

The details pane will slide out showing the details for the default recommendation, and the adjusted recommendation. Perform a what-if analysis by making changes to the adjusted recommendation based on the usage in the graphs, click Apply and see the results.

You can enter the number of vCPUs (in whole numbers) or Gibibytes of memory into the text boxes and click Apply . This will dynamically show the results at that level of commitment.

Hover over the x-axis labels to increase the visibility of each type of usage.

GCP CUDs FAQ

How far back can I go in historical usage for analysis?

A maximum of 2 months: The entire previous months usage, up until the previous day of the current month.

What dimensions can I use to filter the usage for analysis?

You can use any configured Cloudability dimension: Tags & Labels, Account Groups, or Business Mappings. Some dimensions may not be available if the data variability is high – for example: “server name”. Only dimensions with a low-moderate amount of variability will be available, for example: “environment: production/staging”, “cost center: abc123”.

How many dimensions can I use to filter my usage for analysis?

You can select up to 10 Cloudability dimensions to apply to your usage.

When will the selected dimensions be applied to my usage?

Dimensions are applied when the next billing data ingestion occurs. It can take up to a maximum of 24 hrs for the dimensions to be applied to your usage, and be available to be used as filters. When GCP processing starts, it will take the currently configured preferences at that point in time, and apply them.

Can I change my selected dimensions multiple times?

Yes, you can change the dimensions multiple times. When the billing data ingestion occurs, it will take the currently configured dimensions and apply them to the data.

Can different users have different dimensions configured for their usage?

No, the dimensions are configured for the entire organization.

How is the usage on the graphs displayed?

The graphs show hourly data, with multiple hours being averaged into a single point if required to fit into the graph.

How is the default recommendation calculated?

This is the propriety Cloudability algorithm which looks to maximize savings while balancing risk.

Are SUDs incorporated into the recommendation?

Yes, if a resource received a SUD previously, the Cloudability recommendation savings will show the increase from the SUD rate, not the on-demand rate.

The SUD discount level changes throughout a month, what SUD discount level is used?

The average of the entire previous months discount is used and applied to the modelled usage for any resource that received a discount.

How are Cloudability 's recommendations different from the GCP recommendations?

There are 5 features that will result in different results from the native GCP recommendations:

  1. Included data: Cloudability includes any resource running for any amount of time, it does not need to be running consistently for 30 days/1 mont
  2. Usage period: selecting a specific usage period
  3. Filtering the data: Cloudability allows you to remove usage from analysis at a very granular level
  4. SUDs: Cloudability incorporates SUD discounts in the calculations if they were previously received
  5. Risk: Cloudability users can analyze different commitment levels based on risk

How are values in the console rounded ?

Values are rounded to the closest whole number.