Monitoring account resource usage
The Watson Studio, Watson Knowledge Catalog, and Watson Machine Learning plans have restrictions or extra costs for certain types of resources, such as processing usage. If you are the IBM Cloud account owner or administrator, you can monitor the resources that cannot be exceeded or that incur extra cost after the set limits.
For Lite plans, you cannot exceed the limits of the plan. You must wait until the start of your next billing month to use resources that are calculated monthly. Alternatively, you can upgrade to a paid plan.
For Watson Studio and Watson Knowledge Catalog paid plans, you are automatically billed extra when:
- Processing exceeds the amount included in the monthly charge
- The number of users exceeds the amount included in the monthly charge
For Watson Machine Learning paid plans, you are automatically billed extra when:
- Processing exceeds the amount included in the monthly charge
- The number of predictions requested by deployed models exceeds the amount included in the monthly charge
To see the costs of your plan, log in IBM Cloud, open your service instance from your IBM Cloud dashboard, and click Plan. To understand the differences between plans, see Offering plans.
How processing usage is calculated
Processing usage is measured in capacity unit hours (CUH). A capacity unit hour is a specific amount of processing capability with a set cost. Different types of processes and different levels of processing power are billed at different rates of capacity units per hour. For example, the hourly rate for a data profiling process is 6 CUH.
Watson Studio and Watson Knowledge Catalog processes are charged by the minute, with a minimum charge of 10 minutes (0.16 hours). Watson Machine Learning processes are charged by the minute with a minimum charge of one minute.
Processing 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 CUH rate for the process.
These examples show the resulting CUH consumed for data profiling processes that have an hourly rate of 6 CUH:
- 5 minutes: 0.16 hours * 6 CUH = 0.96 CUH. The minimum charge applies.
- 30 minutes: 0.5 hours * 6 CUH = 3 CUH
- 60 minutes: 6 CUH
Processes that consume CUH
Some types of processes, such as running data flows and profiling, have a singe CUH rate. However, runtime environments for processes such as running notebooks, modeler flows, or deep learning experiments, have multiple rates that reflect the memory and processing power for the environment. Environments with more memory and processing power consume capacity unit hours at a higher rate.
These processes consume capacity units:
- Running notebooks: The amount of capacity unit hours that are consumed depends on the runtime environment you choose when you create the notebook. A notebook is running when it is open in edit mode or during a scheduled job. After an hour of inactivity, an environment for a running notebook is stopped. If you choose a Spark service to run your notebook, your compute time is billed to your Spark service.
- Running Spark modeler flows: The amount of capacity unit hours that are consumed depends on the runtime environment you choose when you create the Spark modeler flow in the Modeler tool. A Spark modeler flow is running when it is open in edit mode in the Flow Editor. After an hour of inactivity, an environment for a running modeler flow is stopped. If you choose a Spark service to train your model, your compute time is billed to your Spark service.
- Running Data Refinery flows: The amount of capacity unit hours that are consumed depends on the runtime environment you choose when you create the job for the Data Refinery flow.
- Working in Data Refinery: Capacity unit hours are consumed while you are creating or editing Data Refinery flows. After an hour of inactivity, an environment for Data Refinery is stopped.
- Sampling data assets: Capacity unit hours are consumed while the data assets are being sampled. You can sample data with the IBM Watson API.
- Profiling data assets: Capacity unit hours are consumed while data assets are being profiled. Profiles are automatically created in catalogs with policy enforcement. You can manually create profiles in projects and catalogs without policy enforcement, and manually update existing profiles in catalogs and projects.
- Running deep learning experiments: The amount of capacity unit hours that are consumed depends on the runtime environment you choose when you create the experiment.
- Requesting predictions from deployed models: Capacity unit hours are consumed when deployed models make API calls to generate predictions.
- Generating solutions: Capacity hours are consumed processing decision optimization problems and generating solutions.
This table shows which services provide capacity unit hours for which processes.
|Process||Watson Studio||Watson Knowledge Catalog||Watson Machine Learning|
|Spark modeler flows||✓|
|Working in Data Refinery||✓||✓|
|Data Refinery flows||✓||✓|
How CUH for Data Refinery is applied
If you have both Watson Studio and Watson Knowledge Catalog, the capacity unit hours you can use for working in Data Refinery and running Data Refinery flows in projects are not cumulative. You use the capacity unit hours of the plan with the highest level of precedence:
- Enterprise or Professional
For example, if you have Watson Studio Lite plan and Watson Knowledge Catalog Professional plan, you use Watson Knowledge Catalog CUH.
If the plan level for both apps is the same, then you use CUH from Watson Knowledge Catalog. For example, if you have Watson Studio Lite plan and Watson Knowledge Catalog Lite plan, you can use up to the limit of Watson Knowledge Catalog CUH only. You cannot switch to using Watson Studio CUH.
Monitor monthly billing
You must be an IBM Cloud account owner or administrator to see resource usage information.
To see the monthly charges, the amount of CUH used, and the number of users for any service instance, go to the Cloud Usage Dashboard. For each instance, click Manage > Billing and Usage > Usage, click View Instances next to service name, and then click View instance next to instance name.
Monitor CUH usage
You can view the machine learning environment runtimes that are currently active in a project, and monitor CUH usage for your machine learning assets from the project Environments page.
To see the total amount of CUH used and remaining for Watson Studio, look at the Environment Runtimes page in your Watson Studio app. Choose Manage > Environment Runtimes. This 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.
To see the amount of CUH used and remaining for Watson Machine Learning, go to the Plan page for your Watson Machine Learning service instance in IBM Cloud. Open your Watson Machine Learning service instance from your IBM Cloud dashboard and click Plan.
This table shows how to view the CUH usage by process.
|Notebooks||Environment Runtimes page|
|Modeler flows||Environment Runtimes page|
|Experiments||Watson Machine Learning service instance plan page|
|Predictions||Watson Machine Learning service instance plan page|
Monitor users per app
For Watson Studio, you can see the number of authorized users on the Authorized Users page in your Watson Studio app. Choose Manage > Authorized users. You adjust the number of authorized users as they change. You are automatically billed for users above the set number in your offering plan.
For Watson Knowledge Catalog, you can see the number of users by choosing Manage > Catalogs and clicking Access control in your Watson Knowledge Catalog app. You are automatically billed for users above the set number in your offering plan.