Datasources

DataSource Graphs

Each DataSource definition can have one or more DataSource graphs associated with it.  These graphs will be displayed on the Devices Page for devices that the DataSource applies to.  There are two types of graphs:

  • Normal Graphs- display data per discovered instance (e.g. per disk on a server)
  • Overview Graphs - (multi-instance DataSources only) display an overview of data across the top ten instances of the DataSource (e.g. across all, or the top 10, disks on a server)


The following sections provide instruction for adding, editing, and deleting Normal and Overview DataSource graphs.

Adding & Editing DataSource Graphs

If you'd like to add additional graphs or edit existing graphs for a DataSource, navigate to the DataSource from Settings | LogicModules | DataSources.  Once you've selected the DataSource, you should see the graph definitions in the navigation pane, grouped by normal and overview graphs (if they exist).  If no graphs exist, you can add them using the 'Add Graph' and 'Add Overview Graph' definitions at the top of the DataSource definition:

Adding & Editing Datasource Graphs

Normal Graphs

Normal graphs describe a single instance, and will be displayed for each monitored instance of a DataSource (e.g. a normal graph would be displayed for every disk on a server):

Normal graphs


Displayed Title

The title to be displayed for the graph.  Tokens can be used in this field.

Graph Name

The name that will be used to select the graph to be displayed on dashboards, etc.  

Vertical Label

The label that will appear on the Y axis of the graph.

Min value

The minimum value to display on the graph.  If this value is not set, the graph will auto-scale to cover the range of monitored data.  Setting a minimum value to scale displayed data may be useful if you'd like to compare one graph to another.  

Max value

The maximum value to display on the graph.  If this value is not set, the graph will auto-scale to cover the range of monitored data.  Setting a maximum value to scale displayed data may be useful if you'd like to compare one graph to another.  

Display Priority

This value determines the display order in relation to other graphs for the same DataSource.  Graphs with a lower display priority will be displayed before graphs with a higher display priority.  If two graphs have the same display priority values, they will be displayed in alphabetical order according to the graph titles.

Time scale

This value determines the time scale displayed for the graph.

Datapoints

Select one or more datapoints from the DataSource definition.  You will have the option to select these datapoints to be displayed as graph lines in the 'Lines' configuration section.  For each datapoint added in this section, you'll need to define a Name that the datapoint will be referred to by (this defaults to the DataSource name) and a Consolidation Function.  The Consolidation Function is used to control how the data is displayed when you view the graph data over a longer time period.  For example, consider a datapoint 'aborted_clients' that returned a value of zero for every sample interval during a day, expect for one sample where it instead returned 100:

  • with a Consolidation function of Max, when you look at the graph over a month-long view including the day described above, the day will have a value of 100 (the max value of all samples in the day)
  • with a Consolidation function of Min, when you look at the graph over a month-long view including the day described above, the day will have a value of 0 (the min value of all samples in the day)
  • with a Consolidation function of Average, when you look at the graph over a month-long view including the day described above, the day will have a value of 0.5 (the average value during the day)

Virtual Datapoints

Create one or more virtual datapoints.  Virtual datapoints enable you to perform a calculation using datapoints added to the 'Datapoints' section.  You can use any valid datapoint expression to create virtual datapoints.  You will have the option to select these datapoints to be displayed as graph lines in the 'Lines' configuration section.

Note: There is no history associated with a virtual datapoint, nor are you able to alert on them.  A virtual datapoint is a calculation made specifically for the graph with which it is associated. If you want to associate history or alerts with these datapoints, you should create complex datapoints for your DataSource (using the same datapoint expressions).  Once you create complex datapoints in a DataSource definition, you'll be able to select it to be graphed from the 'Datapoints' section.

Lines

Select the datapoints and/or virtual datapoints you'd like to be displayed as graph lines.  For each line you can select the datapoint to graph (normal or virtual), the type of line, the color of the line and the legend for the line.  As a best practice you should use the color Red for bad conditions and green for good conditions. (e.g. use an area of Green for total disk space, and an area of Red in front for Used disk space.)

Line Types & Legends

Line types include:

  • line
  • area
  • stack
  • column

Note: Stack and Area line types are displayed independently of one another. This means only designated Stack lines can be stacked on top of one another. Stack lines will not be stacked on top of Area lines.

Example: let's say you have the following three lines displayed on a graph:

  • L1= area
  • L2= stack
  • L3= stack

L3 will be stacked on top of L2, but neither will stack on top of L1. Rather, they will simply overlay L1.

Line legends can be static content, or can be populated using dynamic tokens (variables), as is typically the case.  The following tokens are available as line legends:

  • ##HOST##: the device associated with the graphed data
  • ##DATASOURCE##: the DataSource name plus instance name e.g. WinVolumeUsage-C:\
  • ##INSTANCE##: only the instance, not the DataSource. e.g. C:\
  • ##DSIDESCRIPTION##: The DataSource instance description, which may be set by the Active Discovery process, or manually on the Instances Tab to more descriptive text (e.g. "C:\ - finance database volume")
  • Arbitrary device properties: any numeric property defined on the host can be displayed in the legend by referencing the property. e.g. ##COMMITLEVEL##

Overview Graphs

Overview graphs describe the top ten instances, and will be displayed per instance group (by default, a monitored device only has one instance group).  If you define multiple instance groups for a device, an overview graph will be displayed for each of the groups.  For example, an overview graph might display information about all monitored disks on a server.  Note that the top ten instances graphed will always be those with the highest values for the first line defined in the 'Lines' section of the graph.

Overview graph definitions

Displayed Title

The title to be displayed for the graph.  Can use ##HOST## token.

Graph Name

The name that will be used to select the graph to be displayed on dashboards, etc.  

Vertical Label

The label that will appear on the Y axis of the graph.

Min value

The minimum value to display on the graph.  If this value is not set, the graph will auto-scale to cover the range of monitored data.  Setting a minimum value to scale displayed data may be useful if you'd like to compare one graph to another.  

Max value

The maximum value to display on the graph.  If this value is not set, the graph will auto-scale to cover the range of monitored data.  Setting a maximum value to scale displayed data may be useful if you'd like to compare one graph to another.  

Display Priority

This value determines the display order in relation to other graphs for the same DataSource.  Graphs with a lower display priority will be displayed before graphs with a higher display priority.  If two graphs have the same display priority values, they will be displayed in alphabetical order according to the graph titles.

Time scale

This value determines the time scale displayed for the graph.

Aggregate

If this option is selected, the aggregate methods defined for the datapoints added in the 'Datapoints' section will be used.  If this option is left un-selected, the aggregate methods defined for the datapoints will be ignored.  The aggregate method for a datapoint determines how data across multiple instances should be aggregated into one value (e.g. all CPU utilization percents may be averaged across all cores for a device).

Datapoints

Select one or more datapoints from the DataSource definition.  You will have the option to select these datapoints to be displayed as graph lines in the 'Lines' configuration section.  For each datapoint added in this section, you'll need to define a Name that the datapoint will be referred to by (this defaults to the DataSource name) and a Consolidation Function.  The Consolidation Function is used to control how the data is displayed when you view the graph data over a longer time period.  For example, consider a datapoint 'aborted_clients' that returned a value of zero for every sample interval during a day, expect for one sample where it instead returned 100:

  • with a Consolidation function of Max, when you look at the graph over a month-long view including the day described above, the day will have a value of 100 (the max value of all samples in the day)
  • with a Consolidation function of Min, when you look at the graph over a month-long view including the day described above, the day will have a value of 0 (the min value of all samples in the day)
  • with a Consolidation function of Average, when you look at the graph over a month-long view including the day described above, the day will have a value of 0.5 (the average value during the day)

Virtual Datapoints

Create one or more virtual datapoints.  Virtual datapoints enable you to perform a calculation using datapoints added to the 'Datapoints' section.  You can use any valid datapoint expression to create virtual datapoints.  You will have the option to select these datapoints to be displayed as graph lines in the 'Lines' configuration section.

Note: There is no history associated with a virtual datapoint, nor are you able to alert on them.  A virtual datapoint is a calculation made specifically for the graph with which it is associated. If you want to associate history or alerts with these datapoints, you should create complex datapoints for your DataSource (using the same datapoint expressions).  Once you create complex datapoints in a DataSource definition, you'll be able to select it to be graphed from the 'Datapoints' section.

Lines

Select the datapoints and/or virtual datapoints you'd like to be displayed as graph lines.  For each line you'll need to select a line type and specify a legend value.

Line Types & Legends

Line types include:

  • line: data will be plotted in a line
  • area: data will be plotted as an area
  • stack: data will be plotted as an area and the area for each 'line' will be stacked on top of one another
  • column: data will be plotted as a column

Note: Stack and Area line types are displayed independently of one another. This means only designated Stack lines can be stacked on top of one another. Stack lines will not be stacked on top of Area lines.

Example: let's say you have the following three lines displayed on a graph:

  • L1= area
  • L2= stack
  • L3= stack

L3 will be stacked on top of L2, but neither will stack on top of L1. Rather, they will simply overlay L1.

Line legends can be static content, or can be populated using dynamic tokens (variables), as is typically the case.  The following tokens are available as line legends:

  • ##HOST##: the device associated with the graphed data
  • ##DATASOURCE##: the DataSource name plus instance name e.g. WinVolumeUsage-C:\
  • ##INSTANCE##: only the instance, not the DataSource. e.g. C:\
  • ##DSIDESCRIPTION##: The DataSource instance description, which may be set by the Active Discovery process, or manually on the Instances Tab to more descriptive text (e.g. "C:\ - finance database volume")
  • Arbitrary device properties: any numeric property defined on the host can be displayed in the legend by referencing the property. e.g. ##COMMITLEVEL##

Note that if you did not select the aggregate option for the overview graph, you should use the ##INSTANCE## token in the legend, as it will be substituted at graph generation time for the name of the instances displayed (since instance data is not being aggregated into one value).

Deleting DataSource Graphs

You can delete a DataSource graph from the graph definition, using the 'delete' option in the bottom left hand corner:

Deleting Datasource Graphs 

Copying DataSource Graphs to Dashboards

DataSource graphs can be added to a Dashboard by selecting the carrot in the top right corner of a graph and selecting "Add to dashboard" from the dropdown menu. 



Once clicking "Add to dashboard", you will have the opportunity to edit the widget's name, description, and to select the dashboard(s) to which it will be added. All normal graphs will be added to a Dashboard as an editable Custom Graph Widget. The same holds true for overview graphs when there is no instance group under the DataSource.