Manage System Data Exclusion Windows

Data Exclusion Windows are used to resolve unexpected behavior that is observed in headroom or performance impact calculations, or both.

About this task

Peaks in storage system statistics can occur due to:

  • Anomalies or unusual events
  • Recurring maintenance during off-hours that fully loads the storage system

Due to the way this data is condensed and used, unexpected headroom and performance impact results can occur.

Excluding the offending data allows a more accurate result to be calculated. There are two ways to exclude data:

  • One-time exclusion period—When the one-time exclusion period value is set, all statistics before this time are ignored. This setting helps resolve the first case above, where a significant one- time peak distorts the results due to reliance on two weeks of data points. This setting is set system-wide for all components.
  • Recurring exclusion period—You can select one or more 4-hour windows to exclude from performance impact tests and headroom calculations. This setting is set system-wide for all components. Recurring exclusion periods are repeating periods of selected weekday or time slot combinations where collected data is ignored for the purposes of compliance and performance impact considerations. The data is still collected and reported, but it is not used in those calculations.

To manage system data exclusion windows:

Steps

  1. Do one of the following:
    • Select the storage system from the all systems dashboard. Click the SG Compliance panel and In the Actions panel, select Exclude Data.
    • Select the storage system and click Storage > Storage Groups, select a storage group with compliance, go to its Compliance tab, and select Exclude Data > System.
  2. Select All components, for suitability or Back-end components only, for headroom.

Results

The dialog that is displayed allows you to view and set the one-time exclusion period and recurring exclusion periods for a selected storage system.

Each bar in the chart represents a utilization score that is calculated for that time slot. The score itself is the utilization for the array component that is most used during that time slot, that is, the "worst performing" of the array components is the component value that determines the overall value returned. The exact type and identifier of the selected component can be seen in the tooltip for a specific bar. The component types that are represented in the bars are:

  • Front End Port
  • Front End Director
  • Back End Port
  • Back End Director
  • RDF Port
  • RDF Director
  • Cache
  • Thin Pool

For storage systems running PowerMaxOS 10 (6079), back-end port, director, and thin pool utilizations are replaced with an overall back-end utilization.

The colors of the bars signify the following states:

Green represents a utilization value that meets the best practice limit.

Red represents a utilization value that exceeds the best practice limit.

Gray represents a weighted response time that is being ignored before the one-time exclusion period.

Light Gray represents a weighted response time that is being ignored before the recurring exclusion period.

No Bar – If no data was collected or calculated during a time slot, there is no bar present.

The panel consists of a chart that is labeled with the component utilization value as the y-axis and the time slot as the x-axis. Each time slot is a four-hour window during which data was collected and a utilization score was calculated. There is also a horizontal line representing the best practice utilization of 100%. All component utilizations have been normalized to 100% of a component-specific best practice threshold. The x-axis is labeled with the dates of the time slots, that is, the dates of the midnight time slots are labeled with that date and other time slots are blank.

The upper right of this panel has a filter which allows you to include all components that are used in utilization calculations or filter for only those components that are used in headroom calculations. This filter can be helpful when headroom values are causing suitability problems in other areas, but those issues are masked by other component utilizations on this chart. The filters are: All components, for suitability (the default selection) and Back-end components only, for headroom. When you select a value, the page is reloaded with data from the server that is filtered according to the selection made. Both charts are updated to reflect this data.

When the user selects a value, the page is reloaded with data from the server, and filtered according to the selection made. The chart is updated to reflect this data.

You can select and set a time slot before which all collected data is ignored. You select the time slot by clicking the bar. Clicking the bar opens a dialog to choose the type of data exclusion to be applied to the selected bar. The dialog has two check boxes to select the type of exclusion to be applied, and information is displayed under each checkbox to explain the relevant action based on the bar selected (see Manage Data Exclusion actions).

Alerts

There is a system alert generated each time that a user changes a one-time exclusion period value or a recurring exclusion period value.