EP 2.0: Memory usage – IBM

You can view and analyze the current and historical memory usage for a pool, and for the systems and the partitions in the pool.

Memory usage is tracked and charged by the minute. It is tracked based on the assignment of memory and virtual persistent memory (PMEM) to the partitions in the pool. It is based on the average memory usage, not the peak memory usage, for the minute. Virtual PMEM assigned to both active and inactive partitions is tracked and charged since it is always in use. Other memory assigned to inactive partitions is not tracked or charged, however the system firmware memory allocated to support inactive partitions is tracked and charged. All system firmware memory usage is tracked and charged. System firmware memory usage is included in the memory usage shown for systems, and it is also shown on the Memory Usage graph legends area as System Other.

Memory usage is tracked by operating system (O/S) type. However, there is only one type of base memory, and only one type of memory-related charge.

All installed memory on Power S1124, Power S1122, Power S1024, Power S1022, Power S924 (9009-42G) and Power S922 (9009-22G) systems is considered base memory. For these systems, memory is not metered, and the memory rate is not displayed.
You can set a filter to determine which partitions in the pool are displayed. The filter also affects the pool usage and system usage that are displayed. You can filter by the following operating system (OS) types:

AIX
IBM i - P10
IBM i - P20
IBM i - P30
RHEL
RHEL CoreOS
SLES
Other Linux
VIOS

The filter shows only those operating systems that have ever been present in the pool.

The graph on the Memory Usage page shows the memory usage for the pool. Select the appropriate option from the Graph Data Type list to view the following data:

The total memory usage for the pool by operating system type in average TB usage. This graph type is displayed by default.
The total memory usage for the pool by operating system type in GB minutes.
The total metered memory usage for the pool in either GB minutes or in credits.
The peak memory usage in GB for the pool.
A line graph of the peak usage by OS type and total memory is shown. A legend to the right of the graph lists all the OS types and the total memory that is included in the graph. By default, all OS types are included. To exclude an OS type or the total memory from the graph, clear the checkbox from the legend. Alternatively, you can select an operating system using the O/S TYPE filter. If you use the O/S TYPE filter, the graph is automatically updated to show data for only the selected operating system along with the total memory.
Upon hovering the mouse on the graph, peak points are visible. For each peak point, you can see the following details: starting time of peak interval for the hour zoom level or time range for the other zoom levels, peak value, and the time when this peak occurred.
Important: The graph time zoom level in minutes is not available for the Peak Usage Graph Data Type.

By default, the graph zoom level is hours, and each bar in the bar graph or point in the line graph shows the pool memory usage or peak memory usage for one hour. You can use Graph Time Zoom to change the graph zoom level to minutes, hours, days, weeks, or months. Note that days, weeks, and months that are shown in the graph are based on GMT, not local time, but their starting and ending times are shown in local time. For example, a day starts at 12:00 am GMT and ends at 12:00 am GMT the next day, but if your time zone is GMT–5, the starting and ending times shown for a day are 7:00 pm to 7:00 pm. The minute-level memory usage data is automatically deleted three months after it was collected. The hour-level memory usage data is automatically deleted a year after it was collected.

By default, the selected time range in the graph is the current day (midnight to the current hour). You can select a different time range by:

Selecting a time range directly from the calendar. You can select a single day by double-clicking the day in the calendar or you can select a range of days by clicking the starting day and ending day in the calendar.
Selecting a time range in the list that is to the right of the calendar.
Double-clicking a bar or point in the graph to select the time range represented by that single bar or point.
Clicking the graph and dragging the selection box to include the bars in the desired time range.

The memory usage totals displayed just above the graph are for the selected time range. The System Other value is the amount of memory that is used by system firmware. Note that System Other is not represented as a bar in the graph.

Below the graph is a table that contains all the partitions in the pool. You can select Systems just below the graph to view a table that contains all the systems in the pool.

By default, the data in the tables reflect the data for the current day (midnight to the current hour). When you select a different time range, you must click Update Table Data just above the graph to update the data in the tables for the selected time range.

Note that the total memory usage values that are shown in the box just above the partitions in the partitions table do not include the amount of system firmware memory that is allocated to support those partitions. However, the total memory usage values that are shown in the box just above the systems in the systems table and the memory usage values that are shown for the systems in the systems table do include the amount of system firmware memory allocated on the systems.

The current memory usage that is shown in the tables is the memory usage as of approximately 3 minutes before the time the Memory Usage page was loaded.

In the partition table, the Partition Name, ID, O/S Type, Current Memory Usage, Current Persistent Memory Usage, System Name, MTMS, and HMCs columns all show the current information for the partition.

The Average Memory Usage, Average Persistent Memory Usage, and Total GB Minutes columns show the values for the selected time range. The average memory usage includes the time that the partition was not activated, and its memory usage was zero. The Allocated Memory and Allocated Persistent Memory columns show the minimum and maximum value that is assigned to the partition over the selected time range. If the minimum and maximum values are the same, a single value is shown. The minimum and maximum allocated memory values do not consider the configured value when a partition was shut down, but do consider the configured virtual PMEM value when a partition was shut down. The Current Memory Usage, Average Memory Usage, Allocated Memory, and Total GB Minutes columns all include the virtual PMEM amounts, while the Current Persistent Memory Usage, Average Persistent Memory Usage, and Allocated Persistent Memory columns all show only the virtual PMEM amounts.

In the system table, the System Name, HMCs, Installed Memory, and Current Memory Usage columns all show the current information for the system. The Average Memory Usage and Total GB Minutes columns show the values for the selected time range. The memory usage percentages display memory usage as a percentage of installed memory. Virtual PMEM is included in all system and pool memory amounts.

You can click the () icon next to some columns in the tables to change the contents of the columns.

If a current memory usage or average memory usage value that is shown in the system table is displayed in a red box, the value exceeds a configured threshold. To configure system thresholds, visit the Settings page for the pool.

Note: The memory usage values that are shown on this page are all rounded for display purposes. Totals shown in the pool graph legends, totals shown in the totals box just above the systems and partitions tables, and totals shown for system memory usage values are all calculated by using unrounded values, and then rounded for display purposes. Therefore, there might be a slight discrepancy in the total values you calculate using the rounded memory usage values shown in the tables and the total values shown.

Add a Tag: You can also group systems and partitions by assigning them to tag groups. Select the checkbox for the systems and partitions for which you want to assign to a tag and click Add a Tag. Note that the Add a Tag control is grayed out until you select at least one checkbox.

To select an existing tag, select the tag from the list of tags that is displayed, and click Apply.
To create a new tag, enter the tag name and click Create New Tag.
To manage tags or view all resources that are assigned to a tag, click Manage Tags.
When a tag is applied, the tag icon column indicates the number of tags that are associated with the system or partition. Click the tag icon associated with the system or partition to display the tags, and optionally remove them.
You can tag a Red Hat OpenShift Container Platform control node partition with the reserved tag Control Plane Node, which stops software metering for the partition. The Control Plane Node tag is available only for RHEL CoreOS partitions and can be added or removed only in the Enterprise Pools 2.0 app.
You can tag an AIX partition running PowerHA SystemMirror with the relevant reserved tag PowerHA Standard or PowerHA Enterprise to enable the partition to use the corresponding base PowerHA SystemMirror for AIX software licenses in the pool and to be metered for usage above base for the pool. The PowerHA Standard and PowerHA Enterprise tags are available only for AIX partitions and can be added or removed only in the Enterprise Pools 2.0 app.
You can tag an IBM i partition running PowerHA SystemMirror with the reserved tag PowerHA IBM i to enable the partition to use the base PowerHA SystemMirror for IBM i software licenses in the pool and to be metered for usage above base for the pool. The PowerHA IBM i tag is available only for IBM i partitions and can be added or removed only in the Enterprise Pools 2.0 app.

To export the pool, system, and partition memory usage for the selected time range to a file in .csv format, click Export CSV.

To export up to 24 hours of minute-level memory usage data for the pool, click Export Granular Data. To enable this button, you must select Minutes in the Graph Time Zoom. You can export data by using one of the two options that are available in the Export Granular Data pop-up window:

Quick Selection – Click the button for the number of hours prior to the current time for which you want to export the minute-level data. For example, to export data for the prior 24 hours, click Last 24 hours.
Advanced Selection – Select the starting date and time and the number of hours for which you want to export the minute-level data. For example, if you select the date as 1 October, the time as 08:00 a.m., and 12 hours, the data from 01 October 08:00 a.m. to 01 October 08:00 p.m. will be exported.

After selecting which minute-level data you want to export, click Export CSV. The data is exported and downloaded as a .zip file that contains multiple .csv files. To export minute-level data for other dates or times, click Export More Data, and specify a different date or time.

Related articles