Agent directive in Jenkins - Step-by-Step Execution

Process Flow - Agent directive

Start Pipeline

↓

Check Agent Directive

↓

Agent Specified

↓

Allocate Agent

↓

Execute Steps

↓

Pipeline Ends

The pipeline starts, checks if an agent is specified, allocates it if yes, or runs on master if no, then executes steps and ends.

Execution Sample

Jenkins

pipeline {
  agent any
  stages {
    stage('Build') {
      steps { echo 'Building...' }
    }
  }
}

This Jenkins pipeline uses the agent directive to run on any available agent and executes a build stage.

Process Table

Step	Action	Agent Directive	Agent Allocated	Pipeline Step Executed	Output
1	Start pipeline	agent any	Not allocated yet	No	Pipeline started
2	Check agent directive	agent any	Allocating any available agent	No	Agent allocation in progress
3	Agent allocated	agent any	Agent #5 allocated	No	Agent ready
4	Execute stage 'Build'	agent any	Agent #5 allocated	Yes	Building...
5	Pipeline ends	agent any	Agent #5 released	No	Pipeline completed

💡 Pipeline ends after all stages complete and agent is released

Status Tracker

Variable	Start	After Step 2	After Step 3	After Step 4	Final
Agent Directive	agent any	agent any	agent any	agent any	agent any
Agent Allocated	None	Allocating	Agent #5	Agent #5	Released
Pipeline Step Executed	No	No	No	Yes	No

Key Moments - 2 Insights

Why does the pipeline allocate an agent before running steps?

Because the 'agent any' directive tells Jenkins to find any available agent before executing steps, as shown in execution_table row 3.

What happens if no agent directive is specified?

The pipeline runs on the master node by default, skipping agent allocation, which differs from the flow shown here.

Visual Quiz - 3 Questions

Look at the execution table, at which step is the agent actually allocated?

A Step 4

B Step 2

C Step 3

D Step 5

💡 Hint

Check the 'Agent Allocated' column in execution_table row 3

According to variable_tracker, what is the state of 'Pipeline Step Executed' after step 3?

A No

B Yes

C Allocating

D Released

💡 Hint

Look at the 'Pipeline Step Executed' row, column 'After Step 3'

If the agent directive was removed, how would the 'Agent Allocated' column change?

A It would show 'Agent #5 allocated' as before

B It would show 'No agent allocated, runs on master'

C It would show 'Allocating any available agent'

D It would show 'Agent released immediately'

💡 Hint

Recall the key moment about default behavior without agent directive

Concept Snapshot

Jenkins 'agent' directive tells where to run pipeline steps.
Syntax: agent any | none | label 'name'
If agent specified, Jenkins allocates it before running steps.
Without agent, runs on master node by default.
Agent is released after pipeline completes.

Full Transcript

This visual execution shows how Jenkins pipeline uses the agent directive. The pipeline starts and checks if an agent is specified. If yes, it allocates an agent before running the steps. Here, 'agent any' means any available agent is allocated. The build stage runs on that agent. After all stages finish, the agent is released and the pipeline ends. If no agent directive is given, the pipeline runs on the master node by default. Variables like 'Agent Allocated' and 'Pipeline Step Executed' track the state changes during execution.

Practice

1. What does the agent directive specify in a Jenkins pipeline?

easy

A. The build triggers for the pipeline

B. The version of Jenkins to use

C. The environment variables for the pipeline

D. Where the pipeline or stage should run

Solution

1. Step 1: Understand the role of agent

   The agent directive tells Jenkins where to run the pipeline or a specific stage.

2. Step 2: Compare other options

   Options A, B, and C relate to Jenkins configuration but not the agent directive.

3. Final Answer:

   Where the pipeline or stage should run -> Option D

4. Quick Check:

   Agent directive = run location [OK]

Hint: Agent means where to run your pipeline or stage [OK]

Common Mistakes:

• Confusing agent with environment variables
• Thinking agent sets build triggers
• Assuming agent defines Jenkins version

2. Which of the following is the correct syntax to specify that a Jenkins pipeline should run on any available agent?

easy

A. agent any

B. agent: any

C. agent = any

D. agent: 'any'

Solution

1. Step 1: Recall correct Jenkins pipeline syntax

   The correct syntax to specify any available agent is agent any without colons or equals.

2. Step 2: Analyze other options

   Options A, B, and D use incorrect punctuation or quotes not valid in declarative pipeline syntax.

3. Final Answer:

   agent any -> Option A

4. Quick Check:

   Correct syntax for any agent = agent any [OK]

Hint: Use 'agent any' without punctuation for any agent [OK]

Common Mistakes:

• Adding colons or equals signs incorrectly
• Using quotes around any
• Confusing declarative syntax with scripted syntax

3. Given this Jenkins pipeline snippet, what agent will the Build stage run on?

pipeline {
  agent none
  stages {
    stage('Build') {
      agent { label 'linux' }
      steps {
        echo 'Building on Linux agent'
      }
    }
  }
}

medium

A. Agent with label 'linux'

B. Any available agent

C. No agent, pipeline will fail

D. Agent with label 'windows'

Solution

1. Step 1: Understand agent none at pipeline level

   The pipeline disables default agent allocation with agent none.

2. Step 2: Check stage-level agent directive

   The Build stage specifies agent { label 'linux' }, so it runs on an agent labeled 'linux'.

3. Final Answer:

   Agent with label 'linux' -> Option A

4. Quick Check:

   Stage agent label 'linux' overrides pipeline none [OK]

Hint: Stage agent label overrides pipeline agent none [OK]

Common Mistakes:

• Assuming pipeline agent none means no agent anywhere
• Confusing label names
• Thinking default agent applies despite agent none

4. Identify the error in this Jenkins pipeline snippet:

pipeline {
  agent {
    label docker
  }
  stages {
    stage('Test') {
      steps {
        echo 'Testing...'
      }
    }
  }
}

medium

A. Incorrect agent block syntax, should be agent { label 'docker' }

B. Missing quotes around label name

C. Agent directive cannot use label

D. Pipeline missing stages block

Solution

1. Step 1: Review agent block syntax

   The label value must be enclosed in quotes, like label 'docker'.

2. Step 2: Identify syntax error

   The snippet has label docker without quotes, causing a syntax error as docker is treated as an undefined variable.

3. Final Answer:

   Missing quotes around label name -> Option B

4. Quick Check:

   Agent block syntax must be exact [OK]

Hint: Agent block must be exactly agent { label 'name' } [OK]

Common Mistakes:

• Misplacing braces or indentation
• Using agent label without braces
• Confusing agent label with docker agent syntax

5. You want a Jenkins pipeline that does not allocate any agent by default but runs the Deploy stage on a Docker container with image nginx:latest. Which pipeline snippet correctly achieves this?

hard

A. pipeline {
agent any
stages {
stage('Deploy') {
agent none
steps {
echo 'Deploying with nginx container'
}
}
}
}

B. pipeline {
agent { docker { image 'nginx:latest' } }
stages {
stage('Deploy') {
steps {
echo 'Deploying with nginx container'
}
}
}
}

C. pipeline {
agent none
stages {
stage('Deploy') {
agent { docker { image 'nginx:latest' } }
steps {
echo 'Deploying with nginx container'
}
}
}
}

D. pipeline {
agent none
stages {
stage('Deploy') {
agent { label 'nginx:latest' }
steps {
echo 'Deploying with nginx container'
}
}
}
}

Solution

1. Step 1: Disable default agent allocation

   Using agent none at pipeline level disables default agent allocation.

2. Step 2: Specify Docker agent for Deploy stage

   The Deploy stage uses agent { docker { image 'nginx:latest' } } to run inside the nginx Docker container.

3. Step 3: Verify other options

   pipeline { agent { docker { image 'nginx:latest' } } stages { stage('Deploy') { steps { echo 'Deploying with nginx container' } } } } runs entire pipeline in Docker, not just Deploy stage. pipeline { agent any stages { stage('Deploy') { agent none steps { echo 'Deploying with nginx container' } } } } wrongly uses agent none inside stage with agent any pipeline. pipeline { agent none stages { stage('Deploy') { agent { label 'nginx:latest' } steps { echo 'Deploying with nginx container' } } } } uses label instead of docker image.

4. Final Answer:

   Pipeline with agent none and stage agent docker image nginx:latest -> Option C

5. Quick Check:

   agent none + stage docker agent = correct [OK]

Hint: Use agent none globally, docker agent per stage [OK]

Common Mistakes:

• Using agent any globally when none is needed
• Misusing label instead of docker image
• Placing agent none inside stage incorrectly