Monitoring performance

You can monitor performance of your DS8000® storage system to evaluate workload processing and response times.

About this task

The Performance page (Monitoring > Performance) in the DS8000 Storage Management GUI displays graphs that track performance indicators, such as I/O operations per second (IOPS), latency, bandwidth, cache read hits, and write delays. You can create performance graphs for the storage system, pools, and I/O ports. You can also compare performance statistics for multiple pools or I/O ports.

You can also export a system performance summary. For more information, see Exporting the system performance summary.

The following preset graphs are useful for tracking workload processing and response times:

System IOPS: The System IOPS graph displays read, write, and total requests in thousand I/O operations per second (KIOPS), averaged over one minute.
System Latency: The System Latency graph displays the response time in milliseconds (ms) for read and write I/O operations, averaged over one minute.
Note: If your storage system is connected to an IBM® Z host, the System Latency graph includes only partial end-to-end response times on a system level for count key data (CKD) volumes. If you need to view system level response times for a storage system that is connected to a IBM Z host, refer to the Resource Measurement Facility on the host.
System Bandwidth: The System Bandwidth graph displays the number of megabytes per second (MBps) for read, write, and total bandwidth, averaged over one minute.
System Cache: The System Cache graph displays two metrics: the percentage of total read I/O operations that were fulfilled from the storage system cache or memory, and the ratio of write I/O operations that were delayed because of write cache limitations. Both metrics are averaged over one minute.

You can select from the following Performance metrics:

System performance metrics

The system performance metrics represent activity between hosts and the storage system.

I/O

IOPS: Displays the number of processed requests in thousand I/O operations per second (KIOPS) for the selected I/O operations (read, write, and total), averaged over one minute.
Latency: Displays the response time in milliseconds (ms) for the selected I/O operations (read, write, or average of both), averaged over one minute.
Note: If your storage system is connected to an IBM Z host, the system latency metrics include only partial end-to-end response times on a system level for count key data (CKD) volumes. If you need to view system level response times for a storage system that is connected to a IBM Z host, refer to the Resource Measurement Facility on the host.
Transfer Size: The number of KBs per I/O operation (KB/operation) for the selected I/O operations (read, write, or average of both), averaged over one minute.

Bandwidth

Bandwidth: Displays the number of MBs per second (MBps) for the selected bandwidth type (read, write, or total), averaged over one minute.

Cache

Caching: Displays the percentage of read hits (read I/O operations that were fulfilled from the system cache or memory) and write delays (I/O operations that were delayed because of write cache space constraints or other conditions) during one minute.

Capacity

Used IBM Z capacity: The amount of usable capacity that is taken up by IBM Z (CKD) data. The metrics include usable, used, used (average), and provisioned data.
Used Open Systems capacity: The amount of usable capacity that is taken up by Open Systems (FB) data. The metrics include usable, used, used (average), and provisioned data.
Used IBM Z capacity by volume type: The amount of usable capacity that is taken up by IBM Z (CKD) data according to volume type. The volume types include Used (Safeguarded Copy), Used (Thin-provisioned Volumes), Used (Standard Volumes), and Used (Overhead). The metrics also include usable and provisioned data for the system.
Used Open Systems capacity by volume type: The amount of usable capacity that is taken up by Open Systems (FB) data according to volume type. The volume types include Used (Safeguarded Copy), Used (Thin-provisioned Volumes), Used (Standard Volumes), and Used (Overhead). The metrics also include usable and provisioned data for the system.
Provisioned capacity for IBM Z: The total capacity of all IBM Z (CKD) volumes in the system. The metrics include usable, maximum provisioned, and provisioned data.
Provisioned capacity for Open Systems: The total capacity of all Open Systems (FB) volumes in the system. The metrics include usable, maximum provisioned, and provisioned data.
Provisioned IBM Z capacity by volume type: The total capacity of IBM Z (CKD) volumes according to volume type. The volume types include Provisioned (Safeguarded Copy), Provisioned (Thin-provisioned Volumes), and Provisioned (Standard Volumes). The metrics also include usable and maximum provisioned data.
Provisioned Open Systems capacity by volume type: The total capacity of all Open Systems (FB) volumes according to volume type. The volume types include Provisioned (Safeguarded Copy), Provisioned (Thin-provisioned Volumes), and Provisioned (Standard Volumes). The metrics also include usable and maximum provisioned data.

Environment

Power: The power usage of the storage system, in watts.
Temperature: The average inlet temperature of the storage system, in degrees Celsius.

Array performance metrics

The array performance metrics represent activity from the device adapters and drives within the storage system.

I/O

Latency: The response time in milliseconds (ms) for the selected I/O operations (read, write, or average of both) on the I/O port, averaged over one minute.
Transfer Size: The number of KBs per I/O operation (KB/operation) for the selected I/O operations (read, write, or average of both) on the I/O port, averaged over one minute.
Bandwidth: The number of MBs per second (MBps) for the selected bandwidth type (read, write, or total) on the I/O port, averaged over one minute.
Utilization: For each array, displays the percentage of the array that is used by I/O operations.

Bandwidth

Back-end bandwidth: The number of MBs per second (MBps) for the selected bandwidth type (read, write, or total) on the I/O port, averaged over one minute.

Utilization

Utilization %: For each array, displays the percentage of the array that is used by I/O operations.

Pool performance metrics

The pool performance metrics represent activity from the device adapters and drives within the storage system.

I/O

Back-End IOPS: The number of requests that were processed on the arrays in the pool, in thousand I/O operations per second (KIOPS) for the selected I/O operation type (read, write, and total), averaged over one minute.
Back-End Latency: The response time in milliseconds (ms) for the selected I/O operations (read, write, or average of both) that were processed on the arrays in the pool, averaged over one minute.
Transfer Size: The number of KBs per I/O operation (KB/operation) for the selected I/O operations (read, write, or average of both) that were processed on the arrays in the pool, averaged over one minute.

Bandwidth

Back-End Bandwidth: The number of MBs per second (MBps) for the selected bandwidth type (read, write, or total) that was processed on the arrays in the selected pools, averaged over one minute.

Easy Tier

Data Activity: If Easy Tier monitoring is enabled for the pool, display the breakdown of capacity in each tier by the amount of activity, in gibibytes (GiB). The levels of activity in the graph include high activity, sequential activity, low activity, inactive, and available. Click the bar graph for a tier to display an area graph of data activity for that tier over time. Click the graph for a single tier to return to the main graph.
Data Movement: If Easy Tier monitoring is enabled for the pool, display the total movement of data by Easy Tier per day in gibibytes (GiB).
Data Policy: If Easy Tier monitoring is enabled for the pool, display a breakdown of data within the pool that is pinned to a tier or managed by Easy Tier. The graph includes the data that is managed by EasyTier, data that is assigned to a tier, data that is being assigned to another tier, and available data. Click the bar graph for a tier to display an area graph of data policy for that tier. Click the graph for a single tier to return to the main graph.

Capacity

Used capacity: The amount of usable capacity that is taken up by data. The metrics include usable, used, used (average), and provisioned data. The data threshold is also included.
Used capacity by volume type: The amount of usable capacity that is taken up by data according to volume type. The volume types include Used (Safeguarded Copy), Used (Thin-provisioned Volumes), Used (Standard Volumes), and Used (Overhead). The metrics also include usable and provisioned data. The data threshold is also included.
Provisioned capacity: The total capacity of all volumes in a pool. The metrics include usable, maximum provisioned, and provisioned data.
Provisioned capacity by volume type: The total capacity of all volumes in a pool according to volume type. The volume types include Provisioned (Safeguarded Copy), Provisioned (Thin-provisioned Volumes), and Provisioned (Standard Volumes). The metrics also include usable and maximum provisioned data.
Overprovisioned ratio: The ratio of provisioned capacity to usable capacity in a pool.

Volume performance metrics

Easy Tier

Data Activity: If Easy Tier monitoring is enabled for the pool that includes the volume, displays the breakdown of capacity in each tier by the amount of activity, in gibibytes (GiB). The levels of activity in the graph include high activity, sequential activity, low activity, and inactive. Click the bar graph for a tier to display an area graph of data activity for that tier over time. Click the graph for a single tier to return to the main graph.
Data Policy: If Easy Tier monitoring is enabled for the pool that includes the volume, displays a breakdown of data within the pool that is pinned to a tier or managed by Easy Tier. The graph includes the data that is managed by EasyTier, data that is assigned to a tierand data that is being assigned to another tier. Click the bar graph for a tier to display an area graph of data policy for that tier. Click the graph for a single tier to return to the main graph.

I/O

IOPS: The number of requests that are processed on the volume, in thousand I/O operations per second (KIOPS) for default I/O operations (read, write, and total). The time averaged for last hour is sampled per minute, for last day (ranging from last 1 hour to 24 hours ago) is sampled every 15 minutes, for last week (ranging from yesterday to one week ago) is sampled every hour.
Latency: Displays the response time in milliseconds (ms) for the default I/O operations (read, write, and average) that is processed on the volume. The time averaged for last hour is sampled per minute, for last day (ranging from last 1 hour to 24 hours ago) is sampled every 15 minutes, for last week (ranging from yesterday to one week ago) is sampled every hour.

Bandwidth

Bandwidth: The number of MBs per second (MBps) for the default bandwidth type (read, write, and total) that is processed on the volume. The time averaged for last hour is sampled per minute, for last day (ranging from last 1 hour to 24 hours ago) is sampled every 15 minutes, for last week (ranging from yesterday to one week ago) is sampled every hour.

Fibre Channel port performance metrics

I/O

IOPS: Displays the number of processed requests in thousand I/O operations per second (KIOPS) for the selected I/O operations (read, write, and total) on the Fibre Channel port, averaged over one minute.
Latency: Displays the response time in milliseconds (ms) for the selected I/O operations (read, write, or average of both) on the Fibre Channel port, averaged over one minute.
Transfer Size: The number of KBs per I/O operation (KB/operation) for the selected I/O operations (read, write, or average of both) on the Fibre Channel port, averaged over one minute.

Bandwidth

Bandwidth: The number of MBs per second (MBps) for the selected bandwidth type (read, write, or total) on the Fibre Channel port, averaged over one minute.

LSS

I/O

IOPS: The number of requests that are processed on the LSS, in thousand I/O operations per second (KIOPS) for the default I/O operations (read, write, and total) averaged over minute.
Latency: Displays the response time in milliseconds (ms) for the default I/O operations (read, write, and average) that is processed on the LSS, averaged over minute.

Bandwidth

Bandwidth: The number of MBs per second (MBps) for the default bandwidth type (read, write, and total) that is processed on the LSS, averaged over minute.

Host

I/O

IOPS: The number of requests that are processed on the host, in thousand I/O operations per second (KIOPS) for the default I/O operations (read, write, and total) averaged over minute.
Latency: Displays the response time in milliseconds (ms) for the default I/O operations (read, write, and average) that is processed on the host, averaged over minute.

Bandwidth

Bandwidth: The number of MBs per second (MBps) for the selected bandwidth type (read, write, and total) that is processed on the host, averaged over one minute.

Procedure

To monitor overall system performance, use the Performance page to select a performance graph from the Favorites menu or define a new performance graph.

Tips:

To view the performance history, drag the timeline control.
To zoom in or out, use the mouse wheel.
To save a graph to the Favorites menu, click the star in the upper right corner.
To compare two graphs, click the split-screen icon on the right and select the two graphs from the Favorites menu.

To define a performance graph, complete the following steps:

Click the handle on the toolbar in the graph list.
The Resources and Metrics pane opens.
Select the resource type from the drop-down list.
The Resources list displays the selected resources.
From the Resources list, select the resources to display in the graph.
From the Metrics list, select the metrics to monitor.

Note: You can select compatible metrics only.
Click the arrow to display a performance graph based on your selections.