s-Dashboard takes streams and views and displays these visually. In order to get the results you want, you need to configure this data, usually at two stages:
You select input for each dashboard separately. This means that you can view the same data in multiple ways in two dashboards, view streams or views from two different schemas, or view different streams from the same schema.
This page covers the following subtopics:
If you are using StreamLab, panels will open automatically with data from StreamLab. You will only need to follow the steps below if you wish to select a different input source.
The first time you do a Select Input, there may be a delay while s-Dashboard retrieves the catalog for s-Server. s-Server's catalog lists all schemas, streams, tables, and views.
If something changes in your s-Server (for example, a new stream is added), you will need to use the Invalidate the Cache command to refresh the catalog*, as described in the topic Invalidating the Cache in this guide.
To select input:
At any time, you can refresh a panel's connection to its input with the Reopen Input command. This command restarts the connection to the stream, view, or table chosen for input.
Each frame gets its input from one stream, view, or table and displays it using the selected panel. There are separate preferences for how the frame reads its input and for whatever panel is selected.
To set input preferences:
Click the Configure Input Preferences Icon The Input Preferences dialog box opens. You can configure input preferences in four main categories:
Once you configure preferences, click Update. You can change preferences at any time.
The nature of streaming data means that it continually updates. This means that dashboards always display a subset of the data emitted from your source, such as "all failed logins from the last minute" or "all buses speeding over the past hour."
We use the term data window* to refer to how s-Dashboard displays data in terms of time. (Some dashboards can only display one row at a time.) The data window can be a fixed number ("20 rows"), a fixed number per category ("20 rows for each department") a fixed time period ("every hour on the hour"), a sliding time period ("all data from the last sixty minutes").
The first thing you want to determine, then, is how you want to configure your data window. (You can always change this by returning to Input Preferences.)
Choose one of the following four options:
|N Rows||A specified number of rows. When you choose N Rows, a dialog box appears that lets you enter a number. Very high numbers may affect performance.|
|N Rows Per Key||s-Dashboard lets you use a column to divide your data into categories. This column is called the key column, and can be a list of cities, departments, countries, regions, products, and so on--any category into which your data can be grouped. N Rows Per Key works the same as N Rows, but lets you enter a number per category. When you choose N Rows Per Key, a dialog box appears that lets you enter a number. Very high numbers may affect performance.|
|Sliding Time Period||All streaming rows have a timestamp called ROWTIME. This means that you can tell s-Dashboard "give me all the rows in the past hour" or "give me all the rows in the past minute" and so on. Again, high numbers may affect performance, so experiment accordingly.|
|Tumbling Time Period||This data window also uses ROWTIME, and lets you tell s-Dashboard "give me data for an hour, then clear data, then give me data for the next hour," at, for example, 2:00 PM, 3:00 PM, 4:00 PM, and so on.|
Some panels let you group data into categories, using a special column called the key column*. A key column has a limited set of values that categorize the values in other columns, such as cities, departments, product lines, and so on. You can enter the name of this key column in Input Preferences. You need to generate this column as part of your data before using it in s-Dashboard. You could, for example, use the Categorize Analytic in StreamLab to set up a stream with multiple categories.
The panels in the View Rows Over a Key Column display one row for each category in the key column. For example, you can use the Bars Over Key panel to group a bar chart. The stream below has a key column called "key" that contains a list of cities. and producing this chart:
The first time you use the Select Input command, s-Dashboard caches the catalog for s-Server. That is, it identifies all schemas, streams, views, and tables. If your development team has added a new schema, stream, table, or view to s-Server, then these objects will not be available in s-Dashboard unless you use the Invalidate Cache command.
The Invalidate Cache command clears the cache of the currently-stored catalog and requests an updated version of the catalog from s-Server.
To do so:
The old catalog is discarded and an updated catalog is retrieved. When you select input, newly added schemas, streams, tables, and views will appear in the Select Input dialog box.
By default, s-Server assumes views are of streaming data, but they can refer to tables. If you know that your view refers to a table, you can use the Treat View as a Table command to ask s-Dashboard to treat the view as a table (and therefore check the server less frequently.
s-Server rotates files according to options set for FILE_ROTATION_TIME, FILE_ROTATION_SIZE, and FILE_ROTATION_RESPECT_ROWTIME.
You can use FILE_ROTATION_TIME to rotate files based on time elapsed since Unix epoch time. You set FILE_ROTATION_TIME as a positive integer with a unit of either milliseconds (ms), seconds (s), minutes (m), hours (h), or days (d). These express intervals of time from 1970-01-01: an interval might be 15 minutes, 1 hour, or 1 day. Files rotate once a row arrives with a ROWTIME that passes the specified interval.
You can use FILE_ROTATION_SIZE to rotate files based on their size. You specify a file size in kilobytes (k), megabytes (m), or gigabytes (g). Expressed as a positive integer followed by a byte measurement: Defaults to 0. That means "don't use file size to trigger file rotation."
When using FILE_ROTATION_SIZE, you can specify that files wait to rotate until all rows with the same ROWTIME have arrived. For example, if you have set FILE_ROTATION_SIZE to 1m (1 megabyte), and a new row arrives that causes the file to go over the 1 megabyte threshold, if FILE_ROTATION_RESPECT_ROWTIME is set to true, s-Server waits until all rows with the same ROWTIME have arrived. That is, s-Server waits until a new row arrives with a different ROWTIME, even if accepting rows with the same ROWTIME causes the file to grow larger than 1 megabyte. To do so, you set FILE_ROTATION_RESPECT_ROWTIME to true (this is the default behavior).