Analyzing Capacity Utilization

Prev

Analyzing Capacity Utilization

The Prediction tab gives a dashboard-style visualization of the historical and future capacity utilization of clusters registered with Uplink. It shows how your storage capacity is utilized over time.

The historical data is used to make predictions about the capacity utilization of the various clusters in the future, along with the confidence interval, to show the percentage of confidence in the prediction. You can:

  • Select the type of storage capacity to display in the dashboard.

  • Drill down to directories and subdirectories in each cluster to view past and predicted capacity utilization values for each subdirectory

Notice

Prediction is only available for clusters running VAST Cluster 4.5 and later.

Prediction_MainScreen_Callouts_720.png

In the dashboard:

Callout

Description

(1)

The current level of visualization shown as a drill-down path from All Clusters to a specific cluster and its subdirectories, for example:   Prediction_LocationPath.png

(2) and (3)

Historical (2) and future (3) capacity utilization in TB.

Each color represents a cluster registered with Uplink or a directory on the selected cluster.

In the graph:

  • Hover over any cluster or directory visualization to display a popup showing the date and time stamp, name of the cluster or directory and the capacity utilization value at that point in time.

  • Click in any date to pin the date. This freezes the data in the popup window, even when hovering over other dates.

(4)

By default, historic capacity utilization at the All Clusters level is shown as a stacked histogram.

Disable the Stacking option to unstack cluster visualization.

(5)

The Cleanup Aware slider. If disabled, cleanups are not taken into account during capacity estimations.

(6)

Select one of the Capacity Type options:

  • Usable to view the usable storage capacity utilization and predictions.

  • Logical to view the logical storage capacity utilization and predictions.

  • Unique to view the unique data that belongs to a directory or subdirectory (the storage space gained if you delete the directory or subdirectory).

(7)

The circular chart provides the following information:

  • At the All Clusters level: Capacity utilization of specific clusters as a proportion of the total utilized capacity of all clusters.

  • When drilled down to a directory: Capacity utilization of subdirectories as a proportion of the total utilized capacity of the parent directory or cluster.

After you select a point of time in the left pane, this area changes to a graph visualizing the range of predicted capacity utilization values for each cluster or directory at that point of time.

Prediction_RangeGraph.png

For example, the top purple bar in the above graph means that the "purple" cluster will have capacity utilization from 393.814GB to 1034TB at the selected point of time.

Note

When drilled down to directories, only the largest 9 to 10 directories are shown. The files and other directories are grouped together under files & other folders.

(8)

Date range sliders.

The prediction is based on the range selected with the sliders. Use the sliders to exclude known anomalies from the prediction.

(9)

A list of clusters (Cluster) or directories (Directory).

Note

The largest 9 to 10 directories are shown. The files and other directories are grouped together under files & other folders.

(10)

Usage lists capacity utilization in slots depending on the selected time range.

For very small ranges, the maximum granularity is three hours. For larger ranges, the granularity goes up to 12 hours as you move the cursor along the cluster or directory history curve or the projection trendline.

(11)

Click Open subfolder to drill down to subdirectories under the current cluster or directory.

To go back, click the location at the path displayed at the top of the page, for example: Prediction_LocationPath.png

(12)

Click Path confidence indicators to view the confidence interval of the projection for future capacity utilization of the current path. The narrower the confidence interval band, the higher the confidence in the accuracy of the prediction.