Skip to main content

Phase Variables

Last updated on

Phase variables are inputs that are shared across all activities in a phase.

What are Phase Variables?​

Phase variables are:

  • Phase-Scoped: Available only in the phase
  • Inherited: Can inherit from global
  • Phase-Specific: Unique to the phase
  • Isolated: Isolated from other phases

Examples​

Following are examples of phase variables:

  • Environment name for the phase ("staging", "prod")
  • Region for a wave ("us-east-1", "eu-west-1")
  • Common change ticket ID for all activities in the "Change Management" phase

Characteristics​

  • Scoped to the phase: All activities within that phase can reference them via expressions
  • Avoid duplication: Allow you to avoid duplicating the same value on each activity
  • Evaluated by expression engine: Evaluated by the same expression engine reused from pipeline service

Conceptual Example​

  • Phase: "Staging Deployment"
  • Phase variable: env = "staging"
  • Activities:
    • "Deploy Service A" uses env to select environment
    • "Run smoke tests" uses env to choose test environment

Defining Phase Variables​

Phase variables are typically mapped from global variables so users provide only the minimum set of inputs at execution time.

Using Phase Variables​

Variable Resolution​

Resolution order:

  1. Phase variables
  2. Global variables
  3. Process defaults

Variable Isolation​

Phase Isolation​

Variables isolated per phase:

  • Phase A: Has its own variables
  • Phase B: Has its own variables
  • No Sharing: Variables don't share
  • Clear Scope: Clear boundaries

Best Practices​

Appropriate Scope​

Use phase variables when:

  • Variable only needed in phase
  • Phase-specific configuration
  • Isolated phase logic
  • Phase-specific defaults

Clear Naming​

Use descriptive names:

  • Recommended: deployment_max_retries
  • Avoid: retries