The Analysis tab displays historical array data, including storage capacity, consumption, and I/O performance trends across all volumes, and replication bandwidth activity across all source and target groups on the array.
Figure 46. Analysis Tab
The Analysis window contains two main panes:
The navigation pane displays a list of all volumes or protection groups on the array from which you can select to view capacity, performance, and replication data.
The navigation pane is divided into three sections: Capacity, Performance, and Replication.
Click a section heading to expose the list of volumes (for Capacity and Performance) or protection groups (for Replication) and view its details in the details pane.
The Capacity and Performance sections display a list of all volumes on the array.
By default, Purity displays the array capacity or performance details for all volumes across the entire array. To analyze the details for a specific volume, select the check box next to the volume. You can select up to five volumes at one time. The names of the selected volumes appear at the top of the details pane. Clickto clear the volume selection and display a list of all volumes again.
The Replication section displays a list of all source and target protection groups on the array.
By default, Purity displays the bandwidth details for all snapshot data transferred from all source groups and all snapshot data transferred to all target groups on the array. To analyze the details for a specific protection group, select the check box next to the protection group. You can select up to five protection groups at one time. The names of the selected protection groups appear at the top of the details pane. Clickto clear the protection group selection and display a list of all protection groups again.
The details pane displays a series of rolling graphs consisting of real-time capacity, performance, and replication metrics; the incoming data appear along the right side of each graph as older numbers drop off the left side.
The curves in each graph are comprised of a series of individual data points. Hover over any part of a graph to display values for a specific point in time. The values that appear in the point-in-time pop-ups are rounded to two decimal places.
Different graphs display different metrics. Furthermore, specifying all or individual volumes or protection groups determine the metrics that appear within a graph.
A FlashArray maintains a rolling one-year history of data. The granularity of the historical data increases with age; older data points are spaced further apart in time than more recent ones.
Figure 48. Analysis - Details Pane - Point-in-Time Pop-Up
The following example displays the performance statistics for
all volumes on the array at precisely
The following example displays the performance statistics for the
same date/time data points as the example above, but for only two selected volumes,
At the bottom of the details pane is a timeline slider bar and Zoom drop-down. By default, the performance graphs display performance metrics for the past 24 hours.
Click the Zoom drop-down arrow in the bottom-right corner of the window to view statistics from as recently as 1 hour to as far back as the past year. Drag the buttons in the timeline slider bar to further narrow the view to a specific range of time.
For example, zoom in to view data as it trends minute by minute by setting the Zoom value to "1 hour" and dragging the buttons of the slider bar to its narrowest possible setting. Alternatively, set the Zoom value to "1 year" to zoom out and get a birds-eye view of trending data over the past year.
Figure 49. Analysis - Details Pane - Timeline Slider Bar and Zoom Drop-Down
The following example displays the replication statistics for
all protection groups on the array, zoomed out to
30 days and then
further narrowed down to the
06/29-07/5 date range.
Click the title bar of any graph to collapse and expand it.
The graphs in the details pane are grouped into the following areas: Capacity, Performance, and Replication.
The Capacity graphs display array-wide space consumption information, including physical storage capacity and the amount of storage occupied by data and metadata.
Figure 50. Analysis - Details Pane - Capacity Graphs
The Array Capacity graph displays the amount of usable physical storage on the array and the amount of storage occupied by data and metadata. The data point fluctuations represent changes in physical storage consumed by a volume.
For example, a volume may experience a spike in storage consumption when more data is being written to it or when other volumes with shared data are eradicated. Conversely, a volume may experience a dip in storage consumption from trimming or from an increased sharing of deduplicated data with other volumes.
In the Array Capacity graph, the point-in-time pop-up displays the following metrics:
Total physical usable space on the array.
Physical space occupied by internal array metadata.
Physical space occupied by deduplicated data, meaning that the space is shared with other volumes and snapshots as a result of data deduplication.
Physical space occupied by data unique to one or more snapshots.
Physical space occupied by volume data that is not shared between volumes, excluding array metadata and snapshots.
Ratio of mapped sectors within a volume versus the amount of physical space the data occupies after data compression and deduplication. The data reduction ratio does not include thin provisioning savings.
The Host Capacity graph displays the provisioned size of all selected volumes. In the Host Capacity graph, the point-in-time pop-up displays the following metrics:
Total provisioned size of all volumes. Represents storage capacity reported to hosts.
The Performance graphs display I/O performance metrics in real time.
Figure 51. Analysis - Details Pane - Performance Graphs
The I/O type specified in the bottom-left corner of the details pane determines the information that appears in the Performance graphs.
I/O types include:
- R & W (stacked)
The R & W (stacked) charts in the IOPS and Bandwidth graphs compare the read and write parts to the whole. The shaded areas of the chart represent the read and write categories. The topmost shaded line represents the consolidated read and write values stacked on top of one another.
- Total & R & W
The Total & R & W charts display both read and write trends, plus read and write trends combined or averaged, depending on the chart type.
The Read charts display read performance trends.
The Write charts display write performance trends.
The Latency graph displays average internal latency times and SAN times. In the Latency graph, the point-in-time pop-up displays the following metrics:
Average time, measured in milliseconds, required to transfer data between the initiator and the array. SAN times are only displayed in graphs of one I/O type, such as Read or Write.
Average time, measured in milliseconds, it takes the array to perform a read I/O operation.
Average time, measured in milliseconds, it takes the array to perform a write I/O operation.
Average time, measured in milliseconds, it takes the array to perform a read and write I/O operation. Calculated as
Average number of queued I/O requests for all volumes.
Average internal read or write latency time plus SAN time, measured in milliseconds. Calculated as
read+SANwhen viewing Read I/O type, and
write+SANwhen viewing Write I/O type.
The IOPS graph displays the number of read, write, or both read and write requests processed per second. In the IOPS graph, the point-in-time pop-up displays the following metrics:
Number of read requests processed per second.
Number of write requests processed per second.
Total number of read and write requests processed per second.
Average Read IO Size
Average read I/O size per request processed. Calculated as
read bandwidth/read IOPS.
Average Write IO Size
Average write I/O size per request processed. Calculated as
write bandwidth/write IOPS.
Average Read+Write IO Size
Average number of all (both read and write) requests processed per second.
(read bandwidth + write bandwidth)/(read IOPS + write IOPS).
The Bandwidth graph displays the number of bytes transferred per second to and from all hosts. The data is counted in its expanded form (as originally written by the hosts) rather than the reduced form stored in the array to truly reflect what is transferred over the network.
Number of bytes read per second.
Number of bytes written per second.
Total number of bytes read and written second.
Note about the Performance Graphs
The Dashboard and Analysis tabs display the same latency, IOPS, and bandwidth performance graphs, but the information is presented differently between the two tabs.
In the Dashboard tab:
The performance graphs are updated once every 30 seconds.
The performance graphs display up to 30 day's worth of historical data.
The Latency graph displays only internal latency times. SAN times are not included.
In the Analysis tab:
The performance graphs are updated once every minute.
The performance graphs display up to one year's worth of historical data.
The performance data can be further dissected by I/O type.
The Latency graph displays both internal latency times and SAN times.
The Replication details pane consists of one graph named Bandwidth.
Figure 52. Analysis - Details Pane - Replication Graph
The replication Bandwidth graph (not to be confused with the volume Bandwidth graph) displays the number of bytes of replication snapshot data transferred over the storage network per second between protection groups at certain points in time.
For source protection groups, the replication bandwidth value represents the number of bytes of snapshot data transmitted (TX) per second from the source array. For target protection groups, the value represents the number of bytes of snapshot data received (RX) per second by the target array.
In the replication Bandwidth graph, the point-in-time pop-up displays the following metrics:
Number of bytes of replication snapshot data received per second by the target array for the selected protection groups.
Number of bytes of replication snapshot data transmitted per second from the source array for the selected protection groups.
RX + TX
Number of bytes of replication snapshot data transmitted and received per second across all protection groups.