Variable Mapping
Variable mapping allows you to map values between different variable scopes, pass data between phases and activities, and transform variable values.
What is Variable Mapping?
Variable mapping enables the flow of variables from global to phase to activity levels, creating an easy flow of variables and ease of use for end users to understand the minimum set of inputs they have to provide.
Variable mapping enables:
- Scope Mapping: Map between scopes (global → phase → activity)
- Input Simplification: Abstract pipeline inputs for users
Variable Flow
The variable mapping flow works as follows:
- Global Variables are defined at the process level (typically 3-5 key inputs)
- Phase Variables get mapped to global variables
- Example: Phase variable
build_targetgets mapped to a global variable
- Example: Phase variable
- Activity Variables (for pipeline activities) get mapped to phase variables
- Phase variables, in turn, get mapped to global variables
- Activity variables map to the pipeline in case of an automated activity
- Activity Outputs can be captured from pipeline executions and used as inputs to later activities
- Outputs from one activity become inputs to subsequent activities
- This enables dynamic workflows where values generated in early phases flow to later phases
This way, there is an easy flow of variables and ease of use for end users to understand the minimum set of inputs they have to provide.
Activity Output Mapping
Activities can capture outputs from pipeline executions and make those values available to other activities. This is useful when you need to:
- Generate a dynamic value (version number, artifact URL) in one pipeline
- Use that value as input to multiple subsequent pipelines
- Maintain data consistency across different phases
Configuring Activity Outputs
In the source activity (the one generating the value), define outputs that capture pipeline output variables:
activity:
id: generate_version
outputs:
RELEASE_VERSION: <+pipeline.stages.version_stage.spec.execution.steps.generate_step.output.outputVariables.VERSION>
Mapping Outputs to Inputs
In the process input configuration, map the source activity's output to the destination activity's input:
processInput:
phases:
- id: deployment
activities:
- id: deploy_activity
inputs:
VERSION: <+phase.generation.activity.generate_version.outputs.RELEASE_VERSION>
Expression Syntax for Activity Outputs
To reference an activity's output in a process input:
<+phase.<PHASE_ID>.activity.<ACTIVITY_ID>.outputs.<OUTPUT_NAME>>
Example:
- Phase ID:
build_phase - Activity ID:
build_app - Output name:
ARTIFACT_URL - Expression:
<+phase.build_phase.activity.build_app.outputs.ARTIFACT_URL>
When using activity outputs, ensure the destination phase depends on the source phase using depends-on. Without this dependency, phases may execute in parallel, and the output value won't be available when needed.
Go to Passing Activity Outputs to explore a complete walkthrough with examples.
Benefits
Simplified Input Requirements
By mapping variables from global → phase → activity:
- Users only need to provide the key inputs (typically global variables)
- The rest are mapped within the process
- Pipeline inputs are abstracted, making it easier for users who just want to execute and understand the inputs in a simple form
Example Flow
For a complex release process with 3 global variables:
- Global Variable 1:
release_version - Global Variable 2:
environment - Global Variable 3:
deployment_strategy
These global variables will in turn be mapped to:
- Phase variables: Each phase can have variables that map to global variables
- Activity variables: Each activity (especially pipeline activities) can have variables that map to phase variables
When executing the release, users only need to provide these 3 global variables, and the system handles the mapping to phase and activity variables automatically.