Partager via


Specify conditions

TFS 2018

You can specify the conditions under which each stage, job, or step runs. By default, a job or stage runs if it doesn't depend on any other job or stage, or if all of the jobs or stages it depends on have completed and succeeded. This includes not only direct dependencies, but their dependencies as well, computed recursively. By default, a step runs if nothing in its job has failed yet and the step immediately preceding it has finished. You can customize this behavior by forcing a stage, job, or step to run even if a previous dependency fails or by specifying a custom condition.

Note

Microsoft Visual Studio Team Foundation Server 2018 and earlier versions have the following differences in naming:

  • Pipelines for build and release are called definitions
  • Runs are called builds
  • Service connections are called service endpoints
  • Stages are called environments
  • Jobs are called phases

YAML isn't supported in TFS.

Note

When you specify your own condition property for a stage / job / step, you overwrite its default condition: succeeded(). This can lead to your stage / job / step running even if the build is cancelled. Make sure you take into account the state of the parent stage / job when writing your own conditions.

Enable a custom condition

If the built-in conditions don't meet your needs, then you can specify custom conditions.

Conditions are written as expressions in YAML pipelines. The agent evaluates the expression beginning with the innermost function and works its way out. The final result is a boolean value that determines if the task, job, or stage should run or not. See the expressions article for a full guide to the syntax.

Do any of your conditions make it possible for the task to run even after the build is canceled by a user? If so, then specify a reasonable value for cancel timeout so that these kinds of tasks have enough time to complete after the user cancels a run.

Pipeline behavior when build is canceled

When a build is canceled, it doesn't mean all its stages, jobs, or steps stop running. The decision depends on the stage, job, or step conditions you specified and at what point of the pipeline's execution you canceled the build.

If your condition doesn't take into account the state of the parent of your stage / job / step, then if the condition evaluates to true, your stage, job, or step will run, even if its parent is canceled. If its parent is skipped, then your stage, job, or step won't run.

Let's look at some examples.

In this pipeline, by default, stage2 depends on stage1 and stage2 has a condition set. stage2 only runs when the source branch is main.

stages:
- stage: stage1
  jobs:
  - job: A
    steps:
      - script: echo 1; sleep 30
- stage: stage2
  condition: eq(variables['Build.SourceBranch'], 'refs/heads/main')
  jobs:
  - job: B
    steps:
      - script: echo 2

If you queue a build on the main branch, and you cancel it while stage1 is running, stage2 will still run, because eq(variables['Build.SourceBranch'], 'refs/heads/main') evaluates to true.

In this pipeline, stage2 depends on stage1. Job B has a condition set for it.

stages:
- stage: stage1
  jobs:
  - job: A
    steps:
      - script: echo 1; sleep 30
- stage: stage2
  jobs:
  - job: B
    condition: eq(variables['Build.SourceBranch'], 'refs/heads/main')
    steps:
      - script: echo 2

If you queue a build on the main branch, and you cancel it while stage1 is running, stage2 won't run, even though it contains a job B whose condition evaluates to true. The reason is because stage2 has the default condition: succeeded(), which evaluates to false when stage1 is canceled. Therefore, stage2 is skipped, and none of its jobs run.

Say you have the following YAML pipeline. Notice that, by default, stage2 depends on stage1 and that script: echo 2 has a condition set for it.

stages:
- stage: stage1
  jobs:
  - job: A
    steps:
      - script: echo 1; sleep 30
- stage: stage2
  jobs:
  - job: B
    steps:
      - script: echo 2
        condition: eq(variables['Build.SourceBranch'], 'refs/heads/main')

If you queue a build on the main branch, and you cancel it while stage1 is running, stage2 won't run, even though it contains a step in job B whose condition evaluates to true. The reason is because stage2 is skipped in response to stage1 being canceled.

To prevent stages, jobs, or steps with conditions from running when a build is canceled, make sure you consider their parent's state when writing the conditions. For more information, see Job status functions.

Examples

Run for the main branch, even if canceled, even if failing

eq(variables['Build.SourceBranch'], 'refs/heads/main')

Run for the main branch, if succeeding

and(succeeded(), eq(variables['Build.SourceBranch'], 'refs/heads/main'))

Run if the branch is not main, if succeeding

and(succeeded(), ne(variables['Build.SourceBranch'], 'refs/heads/main'))

Run for user topic branches, if succeeding

and(succeeded(), startsWith(variables['Build.SourceBranch'], 'refs/heads/users/'))

Run for continuous integration (CI) builds if succeeding

and(succeeded(), in(variables['Build.Reason'], 'IndividualCI', 'BatchedCI'))

Run if the build is run by a branch policy for a pull request, if failing

and(failed(), eq(variables['Build.Reason'], 'PullRequest'))

Run if the build is scheduled, even if failing, even if canceled

eq(variables['Build.Reason'], 'Schedule')

Release.Artifacts.{artifact-alias}.SourceBranch is equivalent to Build.SourceBranch.

Always run if a variable is set to true, even if canceled, even if failing

eq(variables['System.debug'], true)

Run if a variable is null (empty string)

Since all variables are treated as strings in Azure Pipelines, an empty string is equivalent to null in this pipeline.

variables:
- name: testEmpty
  value: ''

jobs:
  - job: A
    steps:
    - script: echo testEmpty is blank
    condition: eq(variables.testEmpty, '')

Use a template parameter as part of a condition

When you declare a parameter in the same pipeline that you have a condition, parameter expansion happens before conditions are considered. In this case, you can embed parameters inside conditions. The script in this YAML file will run because parameters.doThing is true.

The condition in the pipeline combines two functions: succeeded() and eq('${{ parameters.doThing }}', true). The succeeded() function checks if the previous step succeeded. The succeeded() function returns true because there was no previous step.

The eq('${{ parameters.doThing }}', true) function checks whether the doThing parameter is equal to true. Since the default value for doThing is true, the condition will return true by default unless a different value gets set in the pipeline.

For more template parameter examples, see Template types & usage.

parameters:
- name: doThing
  default: true
  type: boolean

steps:
- script: echo I did a thing
  condition: ${{ eq(parameters.doThing, true) }}

When you pass a parameter to a template, you need to set the parameter's value in your template or use templateContext to pass properties to templates.

# parameters.yml
parameters:
- name: doThing
  default: true # value passed to the condition
  type: boolean

jobs:
  - job: B
    steps:
    - script: echo I did a thing
    condition: ${{ eq(parameters.doThing, true) }}
# azure-pipeline.yml
parameters:
- name: doThing
  default: true 
  type: boolean

trigger:
- none

extends:
  template: parameters.yml

The output of this pipeline is I did a thing because the parameter doThing is true.

Use the output variable from a job in a condition in a subsequent job

You can make a variable available to future jobs and specify it in a condition. Variables available to future jobs must be marked as multi-job output variables using isOutput=true.

jobs:
- job: Foo
  steps:
  - bash: |
      echo "This is job Foo."
      echo "##vso[task.setvariable variable=doThing;isOutput=true]Yes" #set variable doThing to Yes
    name: DetermineResult
- job: Bar
  dependsOn: Foo
  condition: eq(dependencies.Foo.outputs['DetermineResult.doThing'], 'Yes') #map doThing and check the value
  steps:
  - script: echo "Job Foo ran and doThing is Yes."

Use the pipeline variable created from a step in a condition in a subsequent step

You can make a variable available to future steps and specify it in a condition. By default, variables created from a step are available to future steps and don't need to be marked as multi-job output variables using isOutput=true.

There are some important things to note regarding the above approach and scoping:

  • Variables created in a step in a job will be scoped to the steps in the same job.
  • Variables created in a step will only be available in subsequent steps as environment variables.
  • Variables created in a step can't be used in the step that defines them.

Below is an example of creating a pipeline variable in a step and using the variable in a subsequent step's condition and script.

steps:

# This step creates a new pipeline variable: doThing. This variable will be available to subsquent steps.
- bash: |
    echo "##vso[task.setvariable variable=doThing]Yes"
  displayName: Step 1

# This step is able to use doThing, so it uses it in its condition
- script: |
    # You can access the variable from Step 1 as an environment variable.
    echo "Value of doThing (as DOTHING env var): $DOTHING."
  displayName: Step 2
  condition: and(succeeded(), eq(variables['doThing'], 'Yes')) # or and(succeeded(), eq(variables.doThing, 'Yes'))

FAQ

How can I trigger a job if a previous job succeeded with issues?

You can use the result of the previous job. For example, in this YAML file, the condition eq(dependencies.A.result,'SucceededWithIssues') allows the job to run because Job A succeeded with issues.

jobs:
- job: A
  displayName: Job A
  continueOnError: true # next job starts even if this one fails
  steps:
  - script: echo Job A ran
  - script: exit 1

- job: B
  dependsOn: A
  condition: eq(dependencies.A.result,'SucceededWithIssues') # targets the result of the previous job 
  displayName: Job B
  steps:
  - script: echo Job B ran

I canceled my build, but it's still running. What's happening?

You'll experience this issue if the condition that's configured in the stage doesn't include a job status check function. To resolve the issue, add a job status check function to the condition. If you cancel a job while it's in the queue, but not running, the entire job is canceled, including all the other stages.

Learn more about a pipeline's behavior when a build is canceled.