

Pipeline YAML Reference

This document is the reference of the YAML grammar used for describing the pipelines of Semaphore 2.0 projects.

The core properties of a Semaphore pipeline configuration file are blocks, which appear only once at the beginning of a YAML file, task, which can appear multiple times, jobs, which can also be repeated, and promotions that is optional and can appear only once.

Each Semaphore pipeline configuration file has a mandatory preface, which consists of properties version, name and agent.

Properties

version

The version property is a string value that shows the version of the pipeline YAML file that will be used.

The list of valid values for version: v1.0.

Example of version usage:

version: v1.0

name in preface

The name property is a Unicode string that assigns a name to a Semaphore pipeline and is optional. However, you should always give descriptive names to your Semaphore pipelines.

Note: The name property can be found in other sections such as defining the name of a job inside a jobs block.

Example of name usage:

name: The name of the Semaphore 2.0 project

agent

The agent property defines the hardware and software environment of the virtual machine that runs your jobs.

Example of agent usage:

agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804

Note: agent can also be defined on a per task basis.

machine

The machine property, which can only be defined under agent, requires two properties named type and os_image.

Example of machine usage:

machine:
  type: e1-standard-2
  os_image: ubuntu1804

type

The type property is intended for selecting the hardware you would like to use for the Virtual Machine of your jobs.

A complete list of valid values for the type is available on the Machine Types page.

Example of type usage:

type: e1-standard-4

os_image

The os_image is an optional property that specifies the operating system image to be used in the virtual machine. If the value is not provided the default for the machine type is used.

These are valid values for os_image:

• ubuntu1804 (reference)
• macos-xcode11 (reference)

The default operating system depends on the type of the machine:

• For the e1-standard-* machine types the default image is ubuntu1804
• For the a1-standard-* machine types the default image is macos-xcode11

Example of os_image usage:

os_image: ubuntu1804

containers

The containers property, which can only be defined under agent, is an array of docker images that will run the jobs in the pipeline.

Example of containers usage:

agent:
  machine:
    type: e1-standard-2
  containers:
    - name: main
      image: semaphoreci/ruby:2.6.1
    - name: db
      image: postgres:9.6
      user: postgres
      env_vars:
        - name: POSTGRES_PASSOWRD
          value: keyboard-cat

Each container entry must define the name and image property. The name of the container is used when linking the containers together, and for defining hostnames in the first container.

The first container runs the jobs' commands, while the rest of the containers are linked via DNS records. The container with name db is registered with a hostname db in the first container.

Other optional parameters of each container definition can be divided into two groups:

1. Docker related parameters that are passed to docker run command that starts the container. More information about them can be found in docker run docs.

   • user - The user that will be used within the container
   • command - The first command to execute within the container. It overrides the command defined in Dockerfile.
   • entrypoint - This specifies what executable to run when the container starts

2. The data that needs to be injected into containers that is either defined directly there in YAML file or is stored in Semaphore secrets

   • env_vars - Environment variables that are injected into the the container. They are defined in the same way as in task definition.
   • secrets - Secrets which hold the data the should be injected into the container. They are defined in the same way as in task definition. Note: currently, only environment variables defined in a secret will be injected into container, the files within the secret will be ignored.

A Preface example

version: v1.0
name: The name of the Semaphore 2.0 project
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804
blocks:
 - name: Inspect Linux environment
   task:
      jobs:
        - name: Print Envrinment variable
          commands:
            - echo $SEMAPHORE_PIPELINE_ID

execution_time_limit

The concept behind the execution_time_limit property is simple: you do not want your jobs to be executed forever and you need to be able to set a time limit for the entire pipeline as well as for individual blocks and jobs.

The execution_time_limit property can be used at the pipeline, block or job scope. In the latter two cases you can use the propery for multiple blocks or jobs.

Note: The pipeline time limit is time limit for the sum of the running times of all the blocks of the pipeline, this also includes the time jobs are waiting for quota. Please keep this in mind when using the execution_time_limit property.

The execution_time_limit property can hold two other properties, which are named hours and minutes, for specifying the execution time limit in hours, and minutes, respectively. You should only use one of these two properties and not both of them at the same time.

The types of the values for both the hours and minutes properties are integers and their values should be greater than or equal to 1.

If the execution_time_limit property is not used at all, the default global value of the execution_time_limit property is 1 hour. Additionally, if a pipeline has only local execution_time_limit properties, the global value of execution_time_limit will still be 1 hour.

An example with a single execution_time_limit

The following pipeline YAML file uses an execution_time_limit property with a pipeline scope:

version: v1.0
name: Using execution_time_limit
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804
execution_time_limit:
  hours: 3

blocks:
  - name: Creating Docker Image
    task:
      jobs:
      - name: Docker build
        commands:
          - checkout
          - sleep 120

  - name: Building executable
    task:
      jobs:
      - name: Building job
        commands:
          - echo 360
          - echo "Building executable"

An example with an execution_time_limit in a block

The following pipeline YAML file uses two execution_time_limit properties in a block:

version: v1.0
name: Using execution_time_limit
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804

blocks:
  - name: Creating Docker Image
    execution_time_limit:
      minutes: 15
    task:
      jobs:
      - name: Docker build
        commands:
          - checkout
          - sleep 120

  - name: Building executable
    execution_time_limit:
      minutes: 10
    task:
      jobs:
      - name: Building job
        commands:
          - echo 360
          - echo "Building executable"

An example with an execution_time_limit at the job level

The following pipeline YAML file uses the execution_time_limit property with a job scope:

version: v1.0
name: Using execution_time_limit on a job
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804

blocks:
  - name: Tests
    task:
      jobs:

      - name: Lint
        commands:
          - checkout
          - make lint

      - name:  Tests
        execution_time_limit:
          minutes: 30
        commands:
          - checkout
          - make test

An example with multiple execution_time_limit

The following pipeline YAML file uses a combination of an execution_time_limit property with a pipeline scope and a single execution_time_limit property with a block scope:

version: v1.0
name: Using execution_time_limit
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804
execution_time_limit:
  hours: 5

blocks:
  - name: Creating Docker Image
    execution_time_limit:
      minutes: 15
    task:
      jobs:
      - name: Docker build
        commands:
          - checkout
          - sleep 120

  - name: Building executable
    task:
      jobs:
      - name: Building job
        commands:
          - checkout
          - echo "Building executable"

In this case the Creating Docker Image block should finish in less than 15 minutes whereas the Building executable block should finish in less than 5 hours, which is the value of the pipeline scope execution_time_limit property.

However, as blocks are currently executed sequentially, the Building executable block should finish in 5 hours minus the time it took the Creating Docker Image block to finish.

fail_fast

The fail_fast property enables you to set a policy for a pipeline when one of its jobs fail.

It can have two sub-properties, stop and cancel.

At least one of them is required. If both are set, stop will be evaluated first.

Both stop and cancel properties require a condition defined with a when, following the Conditions DSL.

If this condition is fulfilled for a given pipeline execution, the appropriate fail-fast policy is activated.

If a stop policy is set, all running jobs will be stopped and all pending ones will be canceled as soon as possible when one job fails.

If a cancel policy is set, when one job fails, the blocks and jobs which have not yet started executing will be canceled, but already running ones will be allowed to finish executing. This will provide you with more data for debugging without using additional resources.

An example of setting fail-fast stop policy

In the following configuration, blocks A and B run in parallel. Block C runs after block B is finished.

When Block A fails, if the workflow was initiated from a non-master branch, the fail fast stop policy will be applied to:

• Stop block B
• Cancel (ie. not run) block C

version: v1.0
name: Setting fail fast stop policy
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804

fail_fast:
  stop:
    when: "branch != 'master'"

blocks:
  - name: Block A
    dependencies: []
    task:
      jobs:
      - name: Job A
        commands:
          - sleep 10
          - failing command
  - name: Block B
    dependencies: []
    task:
      jobs:
      - name: Job B
        commands:
          - sleep 60
  - name: Block C
    dependencies: ["Block B"]
    task:
      jobs:
      - name: Job C
        commands:
          - sleep 60

An example of setting fail-fast cancel policy

In the following configuration, blocks A and B run in parallel. Block C runs after block B is finished.

When Block A fails, if the workflow was initiated from a non-master branch, the fail fast cancel policy will be applied to:

• Let block B finish
• Cancel (ie. not run) block C

version: v1.0
name: Setting fail fast cancel policy
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804

fail_fast:
  cancel:
    when: "branch != 'master'"

blocks:
  - name: Block A
    dependencies: []
    task:
      jobs:
      - name: Job A
        commands:
          - sleep 10
          - failing command
  - name: Block B
    dependencies: []
    task:
      jobs:
      - name: Job B
        commands:
          - sleep 60
  - name: Block C
    dependencies: ["Block B"]
    task:
      jobs:
      - name: Job C
        commands:
          - sleep 60

queue

The optional queue property enables you to assign the pipeline to the custom execution queues and/or to configure the way the pipelines are processed when queuing is about to happen.

There are two ways you can define a queue behaviour.

The first one is with direct queue configuration that will be applied to all pipelines initiated with a given YAML configuration file.

The second approach is conditional queue configurations that allows you to define an array of queue definitions and conditions under which those definitions should be applied.

All the sub-properties with their possible values for both approaches are listed below and you can find more examples and use cases for different queue configurations on the Pipeline Queues page

direct queue configuration

When using this approach, you can use the name, scope and processing properties as a direct sub-properties of a queue property.

The name property should hold the string that uniquely identifies wanted queue within the configured scope.

If it is omitted, it will be auto-generated based on the git branch/tag name or pull request number and the YAML configuration file name for the given pipeline.

The scope property can have one of two values, project or organization.

If the scope is set to organization the pipelines from the queue will be queued together with pipelines from other projects within the organization that have the queue configuration with same name and scope values.

On the other hand, the queues with the same values for name property in different projects that have scope set to project are mutually independent and their pipelines will not be queued together.

If scope property is omitted, its value will be automatically set to project.

The processing property configures the way pipelines are processed in the queue and it can have one of two values, serialized or parallel.

If the processing is set to serialized the pipelines in the queue will be queued and executed one by one in the ascending order by creation time.

If the processing is set to parallel all pipelines in the queue will be executed as soon as they are created and there will be no queuing at all.

If processing property is omitted, its value will be automatically set to serialized.

Either name or processing property are required for queue definitions to be valid and scope property can only be configured if the name property is also configured.

conditional queue configurations

In this approach, you should define an array of items with queue configurations as a sub-property of the queue property.

Each array item can have the same properties, name, scope and processing, as in direct queue configuration.

Besides those, one additional property is required in each array item and that is a when property that should hold the condition written in Conditions DSL.

When the queue configuration is evaluated in this approach, the when conditions from the items in the array are evaluated one by one starting with the first item in the array.

The evaluation is stopped as soon as one of the when conditions are evaluated as true and the rest of the properties from the same array item are used to configure the queue for the given pipeline.

This means that the order of the items in the array is important and that the items should be ordered so the ones with most specific conditions are defined first followed by the ones with the more generalized conditions (e.g. the one with branch = 'develop' before the one with branch != 'master').

If none of the conditions is evaluated as true, the default queue behaviour will be used.

auto_cancel

The auto_cancel property enables you to set a strategy for auto-canceling other pipelines in a queue when a new one appears.

It can have two sub-properties, running and queued.

At least one of them is required. If both are set, running will be evaluated first.

Both running and queued properties require a condition defined with a when, following the Conditions DSL.

If this condition is fulfilled for a given pipeline execution, the appropriate auto-cancel strategy will be implemented.

If a running auto-cancel strategy is set, the newest pipeline in a queue will:

• cancel all queued pipelines in a queue before self
• stop any running pipeline in a queue before self

This will make new pipelines start running as soon as possible while still making sure that pipelines from the same queue are not run in parallel.

If a queued auto-cancel strategy is set, the newest pipeline in a queue will:

• cancel all queued pipelines in a queue before self
• wait for any running pipeline in a queue before self to finish

This way you will always get results for anything that was already using resources but the new pipelines will be delayed a bit more in a queue.

An example of setting auto-cancel running strategy

In the following configuration, all pipelines initiated from a non-master branch will run immediately after auto stopping everything else in front of them in the queue.

version: v1.0
name: Setting auto-cancel running strategy
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804

auto_cancel:
  running:
    when: "branch != 'master'"

blocks:
  - name: Unit tests
    task:
      jobs:
      - name: Unit tests job
        commands:
          - echo Running unit test

An example of setting auto-cancel queued strategy

In the following configuration, all pipelines initiated from a non-master branch will cancel any queued pipelines and wait for running one to finish before starting to run.

version: v1.0
name: Setting auto-cancel queued strategy
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804

auto_cancel:
  queued:
    when: "branch != 'master'"

blocks:
  - name: Unit tests
    task:
      jobs:
      - name: Unit tests job
        commands:
          - echo Running unit test

global_job_config

The global_job_config property enables you to choose a set of configuration that is shared across the whole pipeline and define it in one place instead of having to repeat it in every task separately.

It can contain any of these properties:

• prologue
• epilogue
• secrets
• env_vars
• priority

The defined configuration values have completely the same syntax as the ones defined on a task or a job level and are applied to all the tasks and jobs in a pipeline.

In the case of prologue and env_vars the global values, ones from global_job_config, are exported first, and after them, the ones defined on a task level. This allows overriding global values for the specific task if there is a need for that.

In the case of epilogue, the order of exporting is reversed, so, for example, one can firstly perform specific cleanup commands before the global ones.

The secrets are just merged since ordering plays no role there.

In the case of the priority, the global values are added at the end of the list of the priorities and their conditions defined on a job level. This allows for job-specific priorities to be evaluated first, and only if none of them matches will the global ones be evaluated and used.

An example of using global_job_config property

version: "v1.0"
name: An example of using global_job_config
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804
global_job_config:
  prologue:
    commands:
      - checkout
  env_vars:
    - name: TEST_ENV_VAR
      value: test_value
blocks:
  - name: Linter
    task:
      jobs:
        - name: Linter
          commands:
            - echo $TEST_ENV_VAR
  - name: Unit tests
    task:
      jobs:
        - name: Unit testing
          commands:
            - echo $TEST_ENV_VAR
  - name: Integration Tests
    task:
      jobs:
        - name: Integration testing
          commands:
            - echo $TEST_ENV_VAR

Blocks

The blocks property defines an array of items that hold the elements of a pipeline. Each element of that array is called a block and can have two properties: name, which is optional and task, which is compulsory.

name in blocks

The name property is a Unicode string that assigns a name to a block and is optional.

If you accidentally name two or more blocks items with the same name value, you will get an error message similar to the following:

semaphore.yml ERROR:

Error: "There are at least two blocks with same name: Build Go project"

Semaphore assigns its own unique names to nameless blocks items, which are displayed in the Semaphore 2.0 UI.

dependencies in blocks

When your pipeline is running blocks in parallel, you can use the dependencies property to define the flow of execution for subsequent blocks.

The following example runs Block A and Block B in parallel at the very beginning of a pipeline.

version: "v1.0"
name: Pipeline with dependencies
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804

blocks:
  - name: "Block A"
    dependencies: []
    task:
      jobs:
      - name: "Job A"
        commands:
          - echo "output"
  - name: "Block B"
    dependencies: []
    task:
      jobs:
      - name: "Job B"
        commands:
          - echo "output"
  - name: "Block C"
    dependencies: ["Block A", "Block B"]
    task:
      jobs:
      - name: "Job C"
        commands:
          - echo "output"

The dependencies property of Block C makes Block A and Block B run in parallel. Once both are finished Block C will run.

If you use dependencies property in some block, you have to specify dependencies for other blocks as well. Otherwise the YAML validation error will be thrown. The following specification is invalid, because dependencies are missing for Block A and Block B.

version: "v1.0"
name: Invalid pipeline
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804

blocks:
  - name: "Block A"
    task:
      jobs:
      - name: "Job A"
        commands:
          - echo "output"
  - name: "Block B"
    task:
      jobs:
      - name: "Job B"
        commands:
          - echo "output"
  - name: "Block C"
    dependencies: ["Block A", "Block B"]
    task:
      jobs:
      - name: "Job C"
        commands:
          - echo "output"

For more examples of common complex CI/CD workflows, see the semaphore-demo-workflows repository on GitHub. You will find specifications for fan-in/fan-out, monorepo and multi-platform pipelines.

task in blocks

All the items of a blocks list have a task property which is required.

You will learn about the properties of a task item in a while.

skip in blocks

The skip property is optional and it allows you to define conditions, written in Conditions DSL, which are based on branch name or tag name of current push which initiated whole pipeline.

If condition defined in this way is evaluated to be true, the block will be skipped.

When block is skipped, it means that it will immediately finish with result passed without actually running any of its jobs.

Its result_reason will be set to skipped and other blocks which depend on it passing will be started and executed as if this block executed regularly and all of its jobs passed.

Note: It is not possible to have both skip and run properties defined for the same block since both of them configure the same behavior, but in opposite ways.

Example of a block that is skipped on all branches except on master:

version: v1.0
name: The name of the Semaphore 2.0 project
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804
blocks:
 - name: Inspect Linux environment
   skip:
     when: "branch != 'master'"
   task:
      jobs:
        - name: Print Environment variables
          commands:
            - echo $SEMAPHORE_PIPELINE_ID
            - echo $HOME

run in blocks

The run property is optional and it allows you to define a condition, written in Conditions DSL, that is based on properties of the push which initiated the whole workflow.

Only if the run condition is evaluated to be true will the block and all of its jobs be run, otherwise the block will be skipped.

When a block is skipped, it means that it will immediately finish with the result passed and the result_reason skipped without actually running any of its jobs.

Note: It is not possible to have both run and skip properties defined for the same block since both of them configure the same behavior, but in opposite ways.

Example of a block that is run only on the master branch:

version: v1.0
name: The name of the Semaphore 2.0 project
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804
blocks:
 - name: Inspect Linux environment
   run:
     when: "branch = 'master'"
   task:
      jobs:
        - name: Print Environment variables
          commands:
            - echo $SEMAPHORE_PIPELINE_ID
            - echo $HOME

Task

The task property that is a compulsory part of each block of a blocks property divides a YAML configuration file into major and distinct sections. Each task item can have multiple jobs items, an optional agent section, an optional prologue section, an optional epilogue section as well as an optional env_vars block for defining environment variables and an optional secrets block for using predefined environment variables from predefined secrets.

jobs

A jobs item is a property of task that allows you to define the commands that you want to execute.

agent in task

The agent section under a task section is optional and can coexist with the global agent definition at the beginning of a Pipeline YAML file. The properties and the possible values of the agent section can be found in the agent reference.

An agent block under a task block overrides the global agent definition.

secrets

The secrets property is used for using existing environment variables from a secret. This is described in the The secrets property section.

prologue

A prologue block is executed before the commands of each job of a task item.

You can consider the prologue commands as a part of each one of the jobs of the same task item.

epilogue

An epilogue block is executed after the commands of each jobs item of a task.

env_vars

The elements of an env_vars array are name and value pairs that hold the name of the environment variable and the value of the environment variable, respectively.

Example of env_vars

version: v1.0
name: A Semaphore 2.0 project
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804
blocks:
   task:
      jobs:
        - name: Check environment variables
          commands:
            - echo $HOME
            - echo $PI
            - echo $VAR1
      env_vars:
           - name: PI
             value: "3.14159"
           - name: VAR1
             value: This is Var 1

The preceding pipeline YAML file defines two environment variables named VAR1 and PI. Both environment variables have string values, which means that numeric values need to be included in double quotes.

Example of task

version: v1.0
name: The name of the Semaphore 2.0 project
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804
blocks:
   task:
      jobs:
        - name: Check environment variables
          commands:
            - echo $SEMAPHORE_PIPELINE_ID
            - echo $HOME
            - echo $PI
            - pwd

      prologue:
        commands:
          - checkout

      env_vars:
        - name: PI
          value: "3.14159"

Caution: The indentation level of prologue, epilogue, env_vars and jobs properties should be the same.

Example of a task block with agent

version: v1.0
name: YAML file example with task and agent.
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804
blocks:
 - name: Run in Linux environment
   task:
      jobs:
        - name: Learn about SEMAPHORE_GIT_DIR
          commands:
            - echo $SEMAPHORE_GIT_DIR

 - name: Run in macOS environment
   task:
      agent:
          machine:
            type: a1-standard-4
            os_image: macos-xcode11
      jobs:
        - name: Using agent job
          commands:
            - echo $PATH

Jobs

The jobs items are essential for each pipeline because they allow you to define the actual commands that you want to execute.

name in jobs

The value of the optional name property is a Unicode string that provides a name to a job.

Semaphore assigns its own names to nameless jobs items, which is what is displayed in the UI.

Tip: It is considered a good practice to give descriptive names to all the jobs and the blocks items of a Semaphore pipeline.

commands

The commands property is an array of strings that holds the commands that will be executed for a job.

Example of commands

The general structure of a job when using the commands property is as follows:

version: v1.0
name: The name of the Semaphore 2.0 project
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804
blocks:
   task:
      jobs:
        - name: Check environment variables
          commands:
            - echo $SEMAPHORE_PIPELINE_ID
            - pwd

commands_file

The commands_file file property allows you to define the path of a plain text file that contains the commands of a job that is an item in a jobs list, a prologue or an epilogue block instead of listing them in a commands list.

You cannot use both commands_file and commands when defining a job, a prologue or an epilogue item. Moreover, you cannot have a job, a prologue or an epilogue properly defined if both the commands and commands_file properties are missing.

Example of commands_file

The contents of the YAML file that defines the pipeline are as follows:

version: v1.0
name: Using commands_file
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804
blocks:
 - name: Calling text file
   task:
      jobs:
        - name: Using command_file
          commands_file: file_with_commands
      prologue:
          commands:
            - checkout

The contents of file_with_commands are as follows:

echo "Command File"
echo "Exit command_file"

Some information about what happens behind the scenes: Semaphore 2.0 reads the plain text file and creates the equivalent job using a commands block, which is what is finally executed. This means that the commands_file property is replaced before the job is started and a machine begins its execution.

env_vars in jobs

An env_vars block can also be defined within a jobs block having a local scope in addition to an env_vars block that is defined on a task level where its scope is the entire task block. In that case the environment variables of that local env_vars block will be only visible to the jobs block it belongs to.

If one or more environment variables are defined on both a jobs level and a task level, the values of the environment variables that are defined on the jobs level take precedence over the values of the environment variables that were also defined on the task level.

Example of env_vars in jobs

version: v1.0
name: Using env_vars per jobs
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804

blocks:
  - name: Using Local Environment variables only
    task:
      jobs:
      - name: Job that uses local env_vars
        commands:
          - echo $APP_ENV
        env_vars:
          - name: APP_ENV
            value: This is APP_ENV
          - name: VAR_2
            value: This is VAR_2 from First Job

  - name: Both local and global env_vars
    task:
      env_vars:
        - name: APP_ENV
          value: prod
        - name: VAR_1
          value: VAR_1 from outer env_vars
      jobs:
      - name: Using both global and local env_vars
        commands:
          - echo $VAR_1
          - echo $VAR_2
          - echo $APP_ENV
        env_vars:
          - name: VAR_1
            value: This is VAR_1
          - name: VAR_2
            value: This is VAR_2
      - name: Second job - no local env_vars
        commands:
          - echo $VAR_1
          - echo $APP_ENV

priority

The priority property allows you to configure a job priority that affects the order in which jobs are started when the parallel jobs quota for the organization is reached.

It holds a list of items where each item has a value property that represents the numerical value for the job priority in range from a 0 to a 100, and a when condition property written in Conditions DSL.

The items are evaluated from the top of the list and the value of the first item which when condition is evaluated as true will be set as the priority for the given job.

If non of the conditions is evaluated as true, the default job priority will be set.

Example of priority

The following pipeline illustrates the use of the priority property:

version: v1.0
name: Job priorities
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804

blocks:
  - name: Tests
    task:
      jobs:
      - name: Unit tests
        priority:
        - value: 70
          when: "branch = 'master'"
        - value: 45
          when: true
        commands:
          - make unit-test
      - name: Integration tests
        priority:
        - value: 58
          when: "branch = 'master'"
        - value: 42
          when: true
        commands:
          - make integration-test

matrix

The matrix property allows you to define one or more environment variable sets with multiple values. In such a setup, n parallel jobs are created, where n equals the cardinality of the Cartesian product of all environment variable sets.

So, the final outcome of the matrix property is the creation of multiple parallel jobs with exactly the same commands that are defined in the respective commands property. Each generated job is assigned with the environment variables from corresponding element of Cartesian product.

Example of matrix

The following pipeline illustrates the use of the matrix property:

version: v1.0
name: Using the matrix property
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804

blocks:
  - name: Elixir + Erlang
    task:
        jobs:
        - name: Elixir + Erlang matrix
          commands:
            - echo $ELIXIR
            - echo $ERLANG
          matrix:
            - env_var: ELIXIR
              values: ["1.3", "1.4"]
            - env_var: ERLANG
              values: ["19", "20", "21"]

In this example, the job specification named Elixir + Erlang matrix expands to 6 parallel jobs as there are 2 x 3 = 6 combinations of the provided environment variables:

• Elixir + Erlang matrix - ELIXIR=1.4, ERLANG=21
• Elixir + Erlang matrix - ELIXIR=1.4, ERLANG=20
• Elixir + Erlang matrix - ELIXIR=1.4, ERLANG=19
• Elixir + Erlang matrix - ELIXIR=1.3, ERLANG=21
• Elixir + Erlang matrix - ELIXIR=1.3, ERLANG=20
• Elixir + Erlang matrix - ELIXIR=1.3, ERLANG=19

parallelism

The parallelism property can be used to easily generate a set of jobs with same commands that can be parameterized. Each of those jobs will have the environment variables with the total number of jobs and the index of a particular job that can be used as parameters.

The parallelism property expects integer value larger than 1.

The following environment variables are added in each generated job:

• SEMAPHORE_JOB_COUNT - total number of jobs generated via parallelism property
• SEMAPHORE_JOB_INDEX - value in the range from 1 to SEMAPHORE_JOB_COUNT which represents the index of a particular job in the list of generated jobs.

Note: It is not possible to have both parallelism and matrix properties defined for the same job since parallelism functionality is a subset of matrix functionality.

Example of using parallelism

When the following configuration is used:

version: v1.0
name: Using the parallelism property
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804

blocks:
  - name: Example for parallelism
    task:
        jobs:
        - name: Parallel job
          parallelism: 4
          commands:
            - echo Job $SEMAPHORE_JOB_INDEX out of $SEMAPHORE_JOB_COUNT
            - make test PARTITION=$SEMAPHORE_JOB_INDEX

It will automatically create 4 jobs with the following names:

• Parallel job - 1/4
• Parallel job - 2/4
• Parallel job - 3/4
• Parallel job - 4/4

Prologue and Epilogue

Each task element can have a single prologue and a single epilogue element. Both prologue and epilogue properties are optional.

The prologue property

A prologue block in a task block is used when you want to execute certain commands prior to the commands of each job of that task. This is usually the case with initialization commands that install software, start or stop services, etc.

Please notice that a pipeline will fail if a command in a prologue block fails to execute for some reason.

Example of prologue

version: v1.0
name: YAML file illustrating the prologue property
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804
blocks:
 - name: Display a file
   task:
      jobs:
        - name: Display hw.go
          commands:
            - ls -al
            - cat hw.go
      prologue:
          commands:
            - checkout

The Epilogue property

An epilogue block should be used when you want to execute some commands after a job has finished, either successfully or unsuccessfully.

Please notice that a pipeline will not fail if one or more commands in the epilogue fail to execute for some reason.

There are three types of epilogue commands:

1. Epilogue commands that are always executed. Defined with always in the epilogue section.

2. Epilogue commands that are executed when the job passes. Defined with on_pass in the epilogue section.

3. Epilogue commands that are executed when the job fails. Defined with on_fail in the epilogue sections.

The order of command execution is the following:

• First, the always commands are executed
• Then the on_pass or on_fail commands are executed

Example of epilogue:

version: v1.0
name: YAML file illustrating the epilogue property
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804
blocks:
 - name: Linux version
   task:
      jobs:
        - name: Execute uname
          commands:
            - uname -a
      epilogue:
        always:
          commands:
            - echo "this command is executed for both passed and failed jobs"
        on_pass:
          commands:
            - echo "This command runs if job has passed"
        on_fail:
          commands:
            - echo "This command runs if job has failed"

Commands can be defined as a list directly in the YAML file, as in the above example, or via a commands_file property:

version: v1.0
name: YAML file illustrating the epilogue property
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804
blocks:
 - name: Linux version
   task:
      jobs:
        - name: Execute uname
          commands:
            - uname -a
      epilogue:
        always:
          commands_file: file_with_epilogue_always_commands.sh
        on_pass:
          commands_file: file_with_epilogue_on_pass_commands.sh
        on_fail:
          commands_file: file_with_epilogue_on_fail_commands.sh

Where the content of the files is a list of commands, like in the following example:

echo "hello from command file"
echo "hello from $SEMAPHORE_GIT_BRANCH/$SEMAPHORE_GIT_SHA"

The location of the file is relative to the pipeline file. For example, if your pipeline file is located in .semaphore/semaphore.yml, the file_with_epilogue_always_commands.sh in the above example is assumed to live in .semaphore/file_with_epilogue_always_commands.sh.

The secrets property

A secret is a place for keeping sensitive information in the form of environment variables and small files. Sharing sensitive data in a secret is both safer and more flexible than storing it using plain text files or environment variables that anyone can access. A secret is defined using a specific YAML grammar and processed using the sem command line tool.

The secrets property is used for importing all the environment variables and files from an existing secret into a Semaphore 2.0 organization.

If one or more names of the environment variables from two or more imported secrets are the same, then the shared environment variables will have the value that was found in the secret that was imported last. The same rule applies to the files in secrets.

Additionally, if you try to use a name value that does not exist, the pipeline will fail to execute.

name in secrets

The name property is compulsory is a secrets block because it specifies the secret that you want to import. Please note that the specified secret or secrets should exist under the active organization.

files in secrets

You can store one or more files in a secret.

You do not need any extra work for using a file that you stored in a secret apart from including the name of the secret that holds the file in the secrets list of the pipeline YAML file. After that the only other requirement is remembering the name of the file, which is the value you put in the path property when creating the secret.

All files in secrets are restored in the home directory of the user of the Virtual Machine (VM), which is /home/semaphore.

Example of secrets with environment variables

version: v1.0
name: Pipeline configuration with secrets
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804
blocks:
  - task:
      jobs:
        - name: Using secrets
          commands:
            - echo $USERNAME
            - echo $PASSWORD
      secrets:
        - name: mysql-secrets

Environment variables imported from a secrets property are used like regular environment variables defined in a env_vars block.

Example of secrets with files

version: v1.0
name: Using files in secrets in Semaphore 2.0
agent:
   machine:
     type: e1-standard-2
     os_image: ubuntu1804

blocks:
- task:
      jobs:
        - name: file.txt
          commands:
            - checkout
            - ls -l ~/file.txt
            - cd ..
            - cat file.txt
            - echo $SECRET_ONE
      secrets:
        - name: more-secrets
        - name: my-secrets

In this case the name of the file that was saved in a secret is file.txt and is located at the home directory of the VM.

promotions

The promotions property is used for promoting (manually or automatically triggering) one or more pipelines using one or more pipeline YAML files, respectively. A pipeline YAML file can have a single promotions block or no promotions block at all.

The items of a promotions block are called targets and are implemented using pairs of name and pipeline_file properties. A promotions block can have multiple targets.

You can promote a target from the UI at any point, even while the pipeline that owns that target is still running.

name in promotions

The name property in a promotions block is compulsory, defines the name of a target and is a Unicode string.

pipeline_file

The pipeline_file property of the promotions block, which is compulsory, is a path to another pipeline YAML file within the GitHub repository of the Semaphore project.

If the pipeline_file value is just a plain filename without any directories, then pipeline_file will look for it inside the .semaphore directory. Otherwise, it will follow the given path beginning from the .semaphore directory.

Each pipeline_file value must be a valid and syntactically correct pipeline YAML file as defined in this document. However, the potential errors of a pipeline YAML file that is given as a value to the pipeline_file property will be revealed when the relevant target is promoted.

The same will happen if the file given as a pipeline_file value does not exist – the error will be revealed at the time of promotion.

Example of promotions

The contents of the .semaphore/semaphore.yml are as follows:

version: v1.0
name: Using promotions
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804

blocks:
  - name: ls
    task:
      jobs:
      - name: List contents
        commands:
          - ls -al
          - echo $SEMAPHORE_PIPELINE_ID

promotions:
  - name: Pipeline 1
    pipeline_file: p1.yml
  - name: Pipeline 2
    pipeline_file: p2.yml

The promotions block in the aforementioned .semaphore/semaphore.yml file will allow you to promote two other YAML files named p1.yml and p2.yml.

The contents of the .semaphore/p1.yml are as follows:

version: v1.0
name: This is Pipeline 1
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804

blocks:
  - name: Environment variable
    task:
      jobs:
      - name: SEMAPHORE_PIPELINE_ID
        commands:
          - echo $SEMAPHORE_PIPELINE_ID

Last, the contents of the .semaphore/p2.yml are as follows:

version: v1.0
name: This is Pipeline 2
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804

blocks:
  - name: List VM Linux version
    task:
      jobs:
      - name: uname
        commands:
          - echo $SEMAPHORE_PIPELINE_ID
          - uname -a

auto_promote

The auto_promote property is optional and it allows you to specify a set of conditions under which the pipeline will be promoted automatically.

It requires conditions to be defined in a when sub-property, following the Conditions DSL.

If these conditions are fulfilled for a given pipeline execution, the appropriate promotion will be triggered automatically.

You can define conditions based on values for the following properties of original pipeline:

• branch - the name of the branch for which pipeline is initiated (empty if it is a tag or pull request)
• tag - the name of the tag for which pipeline is initiated (empty if it is a branch or pull-requests)
• pull request - the number of pull request for which pipeline is initiated (empty if it is a branch or tag)
• result - the result of pipeline's execution, see possible values below
• result_reason - the reason for specific pipeline execution result, see possible values for each result type below

The valid values for result are:

• passed: all the blocks in the pipeline ended successfully
• stopped: the pipeline was stopped either by the user or by the system
• canceled: the pipeline was canceled either by the user or by the system. The difference between canceled and stopped is that if the result is canceled it means that pipeline was terminated before any block or job has started to execute.
• failed: the pipeline failed either due to a pipeline YAML syntax error or because at least one of the blocks of the pipeline failed due to a command not being successfully executed.

The valid values for result_reason are:

• test: one or more of user tests failed
• malformed: the pipeline YAML file is not correct
• stuck: the pipeline was stuck for some internal reason and then aborted
• internal: the pipeline was terminated for internal reasons
• user: the pipeline was stopped on user request
• strategy: the pipeline was terminated due to auto-cancel strategy
• timeout: the pipeline exceeded the execution time limit

Not all result and result_reason combinations make sense. For example, you cannot have passed as the value of result and malformed as the value of result_reason. On the other hand, you can have failed as the value of result and malformed as the value of result_reason.

For a result value of failed, the valid values of result_reason are test, malformed and stuck. When the result value is stopped or canceled, the list of valid values for result_reason are internal, user, strategy and timeout.

Example of auto_promote

The following pipeline YAML file presents an example use of auto_promote and depends on two other pipeline YAML files named p1.yml and p2.yml:

version: v1.0
name: Testing Auto Promoting
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804

promotions:
- name: Staging
  pipeline_file: p1.yml
  auto_promote:
    when: "result = 'passed' and (branch = 'master' or tag =~ '^v1\.')"
- name: Production
  pipeline_file: p2.yml

blocks:
  - name: Block 1
    task:
      jobs:
        - name: Job 1 - Block 1
          commands:
            - echo $SEMAPHORE_GIT_BRANCH

  - name: Block 2
    task:
      jobs:
        - name: Job 1 - Block 2
          commands:
            - echo Job 1 - Block 2
            - echo $SEMAPHORE_GIT_BRANCH
        - name: Job 2 - Block 2
          commands:
            - echo Job 2 - Block 2

According to the specified rules, only the Staging promotion of the promotions list can be auto promoted – in case condition specified in when sub-property of auto_promote property is fulfilled. However, the Production promotion of the promotions list has no auto_promote property so there is no way it can be auto promoted.

So, if the pipeline finishes with result passed and it was initiated from the master branch then the p1.yml pipeline file will be auto-promoted. The same will happen if the pipeline was initiated from the tag that has a name which matches the expression given in PCRE (Perl Compatible Regular Expression) syntax, which is, in this case, any string that starts with v1..

The content of p1.yml is as follows:

version: v1.0
name: Pipeline 1
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804

blocks:
  - name: Environment variable
    task:
      jobs:
      - name: SEMAPHORE_PIPELINE_ID
        commands:
          - echo $SEMAPHORE_PIPELINE_ID

The content of p2.yml is the following:

version: v1.0
name: This is Pipeline 2
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804

blocks:
  - name: List VM Linux version
    task:
      jobs:
      - name: uname
        commands:
          - echo $SEMAPHORE_PIPELINE_ID
          - uname -a

Both p1.yml and p2.yml are perfectly correct pipeline YAML files that could have been used as semaphore.yml files.

auto_promote_on - DEPRECATED

Note: The auto_promote_on property is deprecated in favor of auto_promote property

The auto_promote_on property is used for automatically promoting one or more branches of promotions blocks according to user specified rules.

The auto_promote_on property is a list of items and supports three properties: result, which is mandatory, branch, which is optional, and result_reason, which is also optional.

For a auto_promote_on branch to execute, the return values of all the used properties of that branch must be true.

result

The value of the result property is a string that is used for matching the status of a pipeline.

The list of valid values for result: passed, stopped, canceled and failed.

• passed: all the blocks in the pipeline ended successfully
• stopped: the pipeline was stopped either by the user or by the system
• canceled: the pipeline was cancelled either by the user or by the system. The difference between canceled and stopped is that a pipeline that is not running can be cancelled but cannot be stopped.
• failed: the pipeline failed either due to a pipeline YAML syntax error or because at least one of the blocks of the pipeline failed due to a command not being successfully executed.

branch

The branch property is a list of items. Its items are regular expressions that Semaphore 2.0 tries to match against the name of the GitHub branch that is used with the pipeline that is being executed. If any of them is a match, then the return value of the branch is true.

The branch property uses Perl Compatible Regular Expressions.

In order for a branch value to match the master branch only and not match names such as this-is-not-master or a-master-alternative, you should use ^master$ as the value of the branch property. The same rule applies for matching words or strings.

In order for a branch value to match branches that begin with dev you should try something like ^dev.

result_reason

The value of the result_reason property is a string that signifies the reason that the value of the result property happened.

The list of valid values for result_reason: test, malformed, stuck, deleted, internal and user.

• test: one or more of user tests failed
• malformed: the pipeline YAML file is not correct
• stuck: the pipeline was stuck for some internal reason and then aborted
• deleted: the pipeline was terminated because the GitHub branch was deleted while the pipeline was running
• internal: the pipeline was terminated for internal reasons
• user: the pipeline was stopped on user request

Not all result and result_reason combinations make sense. For example, you cannot have passed as the value of result and malformed as the value of result_reason. On the other hand, you can have failed as the value of result and malformed as the value of result_reason.

For a result value of failed, the valid values of result_reason are test, malformed and stuck. When the result value is stopped or canceled, the list of valid values for result_reason are deleted, internal and user.

Example of auto_promote_on - DEPRECATED

Note: The auto_promote_on property is deprecated in favor of auto_promote property

The following pipeline YAML file presents an example use of auto_promote_on and depends on two other pipeline YAML files named p1.yml and p2.yml:

version: v1.0
name: Testing Auto Promoting
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804

promotions:
- name: Staging
  pipeline_file: p1.yml
  auto_promote_on:
    - result: passed
      branch:
        - "master"
        - ^refs/tags/v1.*
    - result: failed
      branch:
        - "v2."
      result_reason: malformed

- name: prod
  pipeline_file: p2.yml

blocks:
  - name: Block 1
    task:
      jobs:
        - name: Job 1 - Block 1
          commands:
            - echo $SEMAPHORE_GIT_BRANCH
        - name: Job 2 - Block 1
          commands:
            - echo Job 2 - Block 1

  - name: Block 2
    task:
      jobs:
        - name: Job 1 - Block 2
          commands:
            - echo Job 1 - Block 2
            - echo $SEMAPHORE_GIT_BRANCH
        - name: Job 2 - Block 2
          commands:
            - echo Job 2 - Block 2

According to the specified rules only the Staging promotion of the promotions list can be auto promoted – this depends on the rules of the two items of the auto_promote_on list. However, the prod promotion of the promotions list has no auto_promote_on property so there is no way it can be auto promoted.

So, if the pipeline finishes with result passed and the GitHub branch name contains the word master in its name then the p1.yml pipeline file will be auto promoted. The same will happen if the the pipeline finishes with result failed, the result_reason is malformed and the GitHub branch name contains the v2 sequence of characters followed by at least one more character because a . character in a Perl Compatible Regular Expression means one or more characters.

The contents of p1.yml are as follows:

version: v1.0
name: Pipeline 1
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804

blocks:
  - name: Environment variable
    task:
      jobs:
      - name: SEMAPHORE_PIPELINE_ID
        commands:
          - echo $SEMAPHORE_PIPELINE_ID

The contents of p2.yml are the following:

version: v1.0
name: This is Pipeline 2
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804

blocks:
  - name: List VM Linux version
    task:
      jobs:
      - name: uname
        commands:
          - echo $SEMAPHORE_PIPELINE_ID
          - uname -a

Both p1.yml and p2.yml are perfectly correct pipeline YAML files that could have been used as semaphore.yml files.

Complete Configuration examples

A complete .semaphore/semaphore.yml example

The following code presents a complete .semaphore/semaphore.yml file:

version: v1.0
name: YAML file example for Go project
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804

blocks:
 - name: Inspect Linux environment
   task:
      jobs:
        - name: Execute hw.go
          commands:
            - echo $SEMAPHORE_PIPELINE_ID
            - echo $HOME
            - echo $SEMAPHORE_GIT_DIR
            - echo $PI
      prologue:
          commands:
            - checkout
      env_vars:
           - name: PI
             value: "3.14159"

 - name: Build Go project
   task:
      jobs:
        - name: Build hw.go
          commands:
            - checkout
            - change-go-version 1.10
            - go build hw.go
            - ./hw
        - name: PATH variable
          commands:
            - echo $PATH
      epilogue:
        always:
          commands:
            - echo "The job finished with $SEMAPHORE_JOB_RESULT"
        on_pass:
          commands:
            - echo "Executed when the SEMAPHORE_JOB_RESULT is passed"
        on_fail:
          commands:
            - echo "Executed when the SEMAPHORE_JOB_RESULT is failed"

A .semaphore/semaphore.yml file that uses secrets

version: v1.0
name: Pipeline configuration with secrets
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804
blocks:
  - task:
      jobs:
        - name: Using secrets
          commands:
            - checkout
            - ls -l .semaphore
            - echo $SEMAPHORE_PIPELINE_ID
            - echo $SECRET_ONE
            - echo $SECRET_TWO
        - name: Using SECRET_TWO
          commands:
            - checkout
            - echo $SECRET_TWO
            - ls -l .semaphore

      secrets:
        - name: mySecrets
        - name: more-mihalis-secrets

A .semaphore/semaphore.yml file without name properties

Although it is allowed to have .semaphore/semaphore.yml files without name properties, it is considered a very bad practice and should be avoided.

However, the following sample .semaphore/semaphore.yml file proves that it can be done:

version: v1.0
name: Basic YAML configuration file example.
agent:
  machine:
    type: e1-standard-2
    os_image: ubuntu1804
blocks:
  - task:
      jobs:
          - commands:
             - echo $SEMAPHORE_PIPELINE_ID
             - echo "Hello World!"

The order of execution

You cannot and you should not make any assumptions about the order the various jobs items of a task are going to be executed. This means that the jobs of a task item might not start in the order of definition.

However, the blocks items of a .semaphore/semaphore.yml file, which are task items, are executed sequentially. This means that if you have two task items on a .semaphore/semaphore.yml file, the second one will begin only when the first one has finished.

Last, the jobs of a block will run in parallel provided that you have the required capacity (boxes) available.

Comments

Lines that begin with # are considered comments and are being ignored by the YAML parser, which is not a Semaphore 2.0 feature but the way YAML files work.

See also

• Secrets YAML reference
• Projects YAML reference
• sem command line tool Reference
• Toolbox reference page
• Machine Types