Creating and Editing Custom Runbooks Using the Runbook Editor

This topic is a high level overview of the workflow for creating and editing a custom runbook. (A separate video tour of the Runbook Editor is available, also.) For detailed instructions for choosing specific nodes and assembling them on the runbook canvas, refer to Adding and Assembling Runbook Nodes. Refer to Runbook Examples for walkthroughs of some sample runbooks.

Click New on the Runbooks page to open the Runbook Editor and begin defining a new runbook. For an existing runbook, click Edit.

The general workflow is to select a runbook node in the palette at the left and move it onto the runbook canvas at right. You can do this by dragging it to the canvas or by double-clicking it on the palette. The palette lists the runbook node categories in approximate order of how they should be assembled in sequence (first a trigger, then a data query, then a logic operation, then an impact or a visualization), but this is broadly advisory, and a complex runbook may have nodes from different categories arranged in myriad ways. Selecting a node on the canvas opens its properties at the right, enabling you to configure that node's behavior in the context of the larger runbook.

Begin defining a new runbook by choosing a trigger on the palette and moving it to the canvas. Each runbook can have a single trigger, which specifies the set of anomalies that are considered for investigation by the runbook.

In most cases, you'll want to add a data query next, to filter the Data Ocean in the context of the trigger. Choose a data query on the palette and move it to the canvas, to the right of the trigger you chose. If you have the Auto-Connect option active (it is active by default), an arrow will be drawn from the trigger to the data query automatically, connecting the two nodes in a sequence. If Auto-Connect is disabled, you'll need to select the trigger and draw the connection to the data query manually.

Each time you place a new node on the canvas, you can edit its properties, though you can do this in the Runbook Editor at any time, not only at placement. Node properties vary according to the node type, and can be as simple as setting the node name or as sophisticated as specifying metrics and thresholds to control the runbook's behavior.

Make sure to include at least one visualization for each runbook. The runbook will execute and complete its investigation without one, but you won't see any of the output unless there's at least one visualization to show it.

The Comment node (under the Miscellaneous category) enables you to annotate the runbook as you deem appropriate. Comment nodes don't execute or affect the runbook output, and you can use as many of them as you want to document the runbook's behavior.

When you're finished, click the Save icon in the tool bar to save the runbook. It will appear in the list of runbooks shown on the Runbooks page.