What is measured for CUH consumption?

Resources, whether measured with capacity unit hours (CUH) or resource units (RU) are consumed for running assets, not for working in tools. That is, there is no consumption charge for defining an experiment in AutoAI, but there is a charge for running the experiment to train the experiment pipelines. Similarly, there is no charge for creating a deployment space or defining a deployment job, but there is a charge for running a deployment job or inferencing against a deployed asset. Assets that run continuously, such as Jupyter notebooks, RStudio assets, and Bash scripts consume resources for as long as they are active.

How compute usage is calculated

Different types of processes and different levels of compute power are billed at different rates of capacity units per hour. For example, the hourly rate for a data profiling process is 6 capacity units.

Compute usage for watsonx.ai Studio is charged by the minute, with a minimum charge of 10 minutes (0.16 hours). Compute usage for watsonx.ai Runtime is charged by the minute with a minimum charge of one minute.

Compute usage is calculated by adding the minimum number of minutes billed for each process plus the number of minutes the process runs beyond the minimum minutes, then multiplying the total by the capacity unit rate for the process.

The following table shows examples of how the billed CUH is calculated.

Processes that consume capacity unit hours

Some types of processes, such as AutoAI and Federated Learning, have a single compute rate for the runtime. However, with many tools you have a choice of compute resources for the runtime. The notebook editor, Data Refinery, SPSS Modeler, and other tools have different rates that reflect the memory and compute power for the environment. Environments with more memory and compute power consume capacity unit hours at a higher rate.

This table shows each process that consumes CUH, where it runs, and against which service CUH is billed, and whether you can choose from more than one environment. Follow the links to view the available CUH rates for each process.

Monitoring compute usage

You can monitor compute usage for all services at the account level. To view the monthly CUH usage for a service, open the service instance from your IBM Cloud dashboard and click Plan.

You can also monitor compute usage in a project on the Environments page on the Manage tab.

To see the total amount of capacity unit hours that are used and that are remaining for watsonx.ai Studio and watsonx.ai Runtime, look at the Environment Runtimes page. From the navigation menu, select Administration > Environment runtimes. The Environment Runtimes page shows details of the CUH used by environments. You can calculate the amount of CUH you use for data flows and profiling by subtracting the amount used by environments from the total amount used.

Monitoring token usage for foundation model inferencing

You can monitor foundation model token usage in a project on the Environments page on the Manage tab.