Performance configuration

Performance data can be collected from resources that are managed through the native interfaces or from resources that are managed by an SMI-S provider (also called CIM agent or CIMOM). Resources that use an SMI-S provider must be SMI-S compliant.

To monitor the performance of resources, you must complete the following tasks:

1. Prepare for performance monitoring.
   For resources that require an SMI-S provider, ensure that the following conditions are met before you add the SMI-S provider to IBM Spectrum Control:

   • The version of the SMI-S provider and the firmware for the device is supported. For information about the SMI-S provider and firmware that is supported, see https://www.ibm.com/support/pages/node/388393 and go to the Storage section.
   • An SMI-S provider is installed on a different server than the IBM Spectrum Control server. For some switch vendors, the switch comes with an embedded SMI-S provider, so an SMI-S provider does not need to be installed.
   • For storage systems or switches on a private network, ensure that the SMI-S provider is installed on a gateway machine so that the IBM Spectrum Control server can communicate with that agent.
   • The SMI-S provider is configured to manage the intended resource.
2. Add resources for monitoring and schedule data collection.
   Before you can view the performance of a resource, you must add it to IBM Spectrum Control for monitoring. When you add resources, you can automatically schedule performance monitors. Performance monitors are data collection jobs that collect performance information.

   For more information about adding resources, see Adding resources and search for adding resources.

3. Use the performance alerts in the default alert policies or create alerts for performance thresholds.
   To be notified when the performance of a resource might represent a potential problem, define a threshold alert to monitor a specific metric. In a threshold alert, you can specify one or more boundary values for a metric. When the performance of a resource falls outside these boundaries, an alert is generated. For example, you can define an alert threshold that notifies you when the total I/O rate for a port on a storage system falls outside a specified range. Chronologically, you must define a threshold alert before a performance monitor starts collecting data.

   Alerts are triggered by conditions and violations of performance thresholds that are detected during data collection and event processing. For some storage systems such as IBM Storage Accelerate and the XIV, events are polled every minute from the resource. For IBM Storage Scale, status change events are polled frequently, typically within minutes. For other resources, events are subscription-based, where the resource itself or a data source such as an SMI-S provider sends the events to IBM Spectrum Control when conditions change on the resource. Examples of storage systems that use subscription-based event processing include SAN Volume Controller, Storwize V7000, Storwize V7000 Unified, and IBM Storage FlashSystem V9000.

   Tip: Alerts can be created, deleted, or modified at any time, even if the performance monitor is already running. Changes are applied dynamically to the running performance monitor and take effect the next time that data is collected.
4. Check the status of a performance monitor.
   When a performance monitor starts to run, IBM Spectrum Control begins to collect performance data for the specified resource. To check the status of the performance monitor to ensure that it is running or completed successfully, complete the following steps:

   1. In the menu bar, go to Home > Performance Monitors.
   2. In the Name column, locate the name of the resource that was included in the performance monitor.
   3. In the Status column, check the status of the performance monitor.
5. Specify how long performance data is retained.
   In the GUI, go to Settings > History Retention to configure how long to retain performance data that IBM Spectrum Control collects about resources. You can specify a retention period for collected sample data, for aggregated hourly data, and for daily data. The retention period applies to all performance monitors. The longer that you keep data, the more informative your analysis.

   Sample data is the data that is collected at the specified interval length of the performance monitor. For example, data is collected every 5 minutes. For storage systems, the most numerous components are usually volumes, therefore the largest amount of performance data is collected for volumes.