Mapping functions
StackState Self-hosted v5.1.x
Last updated
StackState v5.1
StackState Self-hosted v5.1.x
Last updated
Mapping functions can optionally be used in topology synchronization to transform data. The goal of a mapping function is to process topology data from an external system and prepare parameters for use by a template function. There are two types of mapping functions:
Component mapping functions - used in the synchronization of components.
Relation mapping functions - used in the synchronization of relations.
To add a custom component or relation mapping function:
In the StackState UI go to the settings page for the type of mapping function you would like to add:
For component mapping functions, go to: Settings > Topology Synchronization > Component Mapping Functions
For relation mapping functions, go to: Settings > Topology Synchronization > Relation Mapping Functions
Click ADD COMPONENT MAPPING FUNCTION or ADD RELATION MAPPING FUNCTION.
Enter the required settings:
Name - The name of the mapping function is shown on the settings page and in the topology synchronization that uses it.
Description - Optional. Describes the mapping function in further detail. Will be shown on the mapping function settings page.
Parameters - Mapping functions work with one system parameter element. This is used to process the component or relation JSON payload. It isn't possible to add additional user parameters.
Script - Defines the logic of the mapping function using the Groovy scripting language.
Identifier - Optional. Specifies the URN identifier for the mapping function. This should be set if you intend to include the mapping function in a StackPack.
Click CREATE to save the mapping function.
The mapping function will be listed on the settings page (Component Mapping Functions or Relation Mapping Functions).
The mapping function is available for use in .
This example shows a simple component mapping function - the Fixed label mapping function - that adds a fixed label to the component’s data payload. This mapping function could also be replaced by adding the appropriate label block directly in the component template that is used.
The script checks if the data payload already has a key named labels, as shown on line 1. If the labels key isn't present, then the value is initialized to an empty list to ensure that the value is going to be a list. If the labels key does exist, the assumption is made that the value is a list.
Line 5 appends the label label-added-by-mapper-function to the list of labels. The updated element variable is returned by the mapping function on line 7.
Mapping functions are specified in step 3/5 of a topology synchronization. Here, the Fixed label mapping function is added to the Other sources mapping. In essence, every component processed by this topology synchronization will receive the label. A generic component template is used that converts the data.labels payload to a StackState label.
The process and resulting component is shown here.
Component JSON - The component JSON object that was sent to the StackState receiver as part of a larger JSON payload is shown on the left. The JSON defines a component with name Application1. The JSON payload is processed by the topology synchronization on which the mapping function was configured.
Mapping function script - A part of the mapping function is shown at the bottom to show the value of the label.
Component template - A part of the component template is shown in the center. In the component template, the labels payload is processed using the join helper, as shown on line 30. This iterates through each string in the label list and outputs the JSON block on lines 31 until 34. The this variable inside the join helper on line 33 is used access the current value in the iteration. In this example, there is only one label to be processed. This is because there is no labels key in the original JSON and the mapping function adds the only one. Lines 37 until 40 also define a fixed label. The label defined there will have the value demo:demo. This shows that, in this example, the mapping function could also be replaced by adding the appropriate label block directly in the component template.
Component properties - On the right is the Component Details tab of the Application1 component after the JSON payload has been processed by topology synchronization. The Application1 component has two labels:
The demo:demo label that was added by the component template.
The label-added-by-mapper-function that was added by the mapping function .
This example is used for hosts reported by StackState Agent. The mapping function adds the information needed to create a metric stream to the data payload, the template can then convert the information to actual metric streams.
The mapping function script is shown below.
Line 1 of the script sets the value of the key streams to be a list.
The entries of that list are also added, as shown on lines 2 through to 14. Each entry will be a map of key:value pairs with enough information about the metric stream for the template to create the stream definition. Each entry has the keys name, metric, id, and priority.
Line 17 has an if statement checking whether the operating system is Linux by checking the os.linux data payload. If the if statement resolves to True, then two more metric streams are added to the streams list. This illustrates that the mapping function can be used to make decisions based on the received data to change the resulting topology element.
The template and a resulting stream are shown here.
From the mapping function, one entry from the streams list is shown at the top to illustrate how the component template handles the data. The streams section of the component template is shown on the left. On the right, the resulting metric stream is shown.
The template shows the streams block starting on line 53 in the screenshot above. For easy reference, the full streams block is shown below.
Line 2 in the streams block below shows the join helper being used to iterate over each entry in the streams list.
Inside the join helper’s code block, a definition of a metric stream is placed. Each entry in the streams list adds a new stream definition to the resulting template JSON. Variables are used to create stream blocks dynamically. These variables are available on each entry of the streams list. Each entry has the following variables: name, metric, id, and priority.
The name key is referenced on line 5 to add the value as the name of the stream. In the screenshot above, the resulting name of the stream is shown on the right.
The conditions block on line 7 adds two conditions for each metric stream.
Lines 8 through to 11 define a filter for the host. The value for that condition is obtained from element.data.host and not from the streams list.
The condition on lines 12 through to 15 references the metric key provided by the streams list entry.
In the screenshot above, the name condition is shown on the right, in the Select drop-down.
After processing, the resulting component has quite some metric stream definitions - one metric stream definition for each entry in the streams list.
The list and the join helper show that definitions can be created dynamically.
Alternatively, it would have been possible to add each metric stream definition inside the streams block, however, this would make the streams block somewhat large and repetitive.
