Defining and testing data sources

Setting up a data source for IBM Cloud Pak for Multicloud Management

In IBM® Cloud Pak for Multicloud Management, you can use data from all data sets in the thresholds that you create. For data to be visible in the IBM Cloud Pak console, you must model the data as one or more resources.

These agent resources should group subsets of the data so that each resource represents a logical entity in the application, system, or network environment. Each resource can contain any subset of the information contained in any number of data sources. Each resource definition should include one data source with at least one attribute that can be used to identify the resource. If the data source is single-row, the agent creates one resource. If the data source is multi-row, the agent creates a resource for each unique set of values. A resource can include an event data source as additional data. All of the data selected when the resource is defined is displayed in a table in the IBM Cloud Pak console. You can choose to plot a subset of the data in a line graph by specifying a units value for the attribute.

For more information, see Preparing the agent for Cloud Pak for Multicloud Management.

Setting up a data source for IBM Cloud Application Performance Management

In Cloud APM, you can use data from all data sets in the Details dashboard and to set up thresholds using the threshold manager. If you want to use information from a data set in the summary dashboard for the agent or subnode, including the status indicator, as well as for resource information (service name, address, and port), the data set must produce only one row.

For most data providers, you can select Produces a single data row in the data set configuration. If the gathered information would include more than one row, you can click Advanced to set up a filter that ensures the correct row is produced (for instructions, see Filtering attribute groups). You can test your data source to ensure that the gathered information produces the row that you need.

For some data providers, the data set must produce multiple rows. Also, the process, Windows service, and command return code data sources place data into a single Availability data set, which produces multiple rows. In such cases, you must create a filtered data set that produces one row. For instructions about creating a filtered data set (attribute group), see Creating a filtered attribute group.

Some other data providers produce event data; a row is included for every new event. Do not use these data providers for summary or resource information in Cloud APM.

One of the attributes of the data set must provide a status value. Cloud APM uses this value for the overall status indicator. If the row does not include an attribute that can be used as a status indicator, you can create a derived attribute to calculate the status. You must configure the status severity values; for instructions, see Specifying severity for an attribute used as a status indicator.