Using the Data Flow Tool

The data displayed in the Data Flow page is sampled at 60 second granularity. It may take up to 60 seconds to see I/O reflected in data-flows.

The Defined Time field at the top left of the Data Flow page enables you to select the sampling time for which to display data flow analytics.

The options available are:

Points on data flow paths (users, hosts, VIPs and so on) are sorted in their columns based on multiple criteria. Only top N performers are displayed, up to a limit set by the Limit field, and with the Protocol filter applied.

For any column, you can toggle between descending and ascending sorting order by clicking the arrow next to the column's name ().

To set sorting and filtering criteria:

You can set a limit on the number of top performers that are shown in each of the columns in the Data Flow page. Data flow points that are not displayed due to the limit can be found through a search.

To set the maximum allowed number of data flow points to show, select a value from the Limit dropdown. The default is 10.

When the number of data flow points in more than one category is greater than the maximum, the number of data flow points shown might sometimes be further reduced from the maximum allowed value in each category. This is because data flow points are only displayed if they share paths with the other displayed data flow points.

Use the slider to the left of the Vip column to toggle between displaying the flow through each virtual IP individually and displaying the flow through each virtual IP pool.

Depending on the current setting, the column name is Vip or Vippool.

Hover over any object on the page, such as a virtual IP pool, to see the IO metrics for this object, averaged over the sampling period.

The IO metrics include average bandwidth, IOPS and metadata IOPS for read and/or write operations, and also statistics on objects along the data flow path, and ROW operations on Databases.

The metrics displayed when you hover over an object reflect the aggregated flows included based on the sample period and the selected filters. They are recalculated when you change the sampling period, sort the flow points, or change the limit. However, they are not recalculated when the object is pinned or the paths through the data flow point are isolated as described in Focusing on Specific Data Flows and Objects.

The search feature searches the sampled IOs for the selected time period, and can find a data flow point and the IOs that went through that point during the sampling period even if the object at the data flow point is a low performer (according to the current sorting and filtering criteria) and is not displayed due to the limit.

If searching does not find an object (such as a CNode or a view), this means that IO did not flow through the object.

To search, enter text into the search field below a column's name. Partial text search is supported and can return multiple results. For example, searching for cnode-6 would also display CNodes named cnode-60, cnode-61, cnode-62, and so on.

The data displayed in the Data Flow page can be refreshed on demand or automatically as follows:

To investigate data flow paths that go through a specific object, click the object to pin it. Pinning an object highlights data flow paths going through it, so that you can focus on these specific paths, while other paths appear greyed out.

You can pin one or more flow points in one column at a time. To unpin a flow point, click it again.

When flow points are pinned, you can isolate all the flows that go through those objects. Isolating flows gives you a simplified view of the paths you want to focus on, while hiding all the paths that do not flow through the pinned objects.

To isolate flows, first pin one or more objects and then click Isolate Selected Flows. When you're done looking at the isolated flows, you can clear them by clicking Clear isolation.