Pipeline Durability

Learn how to configure and optimize Jenkins Pipeline Durability to balance performance, disk I/O, and recoverability. This guide covers durability levels, configuration scopes, and a hands-on demonstration.

What Is Pipeline Durability?

Pipeline durability determines how frequently Jenkins writes the pipeline state to disk. Frequent writes ensure resumability after crashes but can introduce latency. Jenkins offers three levels:

Where to Configure Durability

You can apply durability settings at three levels:

1. Global – Default for all jobs.
2. Multibranch project – Per-branch overrides.
3. Single pipeline job – Specific to one Declarative Pipeline.

1. Global Configuration

1. Go to Manage Jenkins → Configure System.
2. Search for Pipeline Speed and Durability.
3. Select your default level.

In newer Jenkins releases, you may also see a dedicated dropdown:

2. Multibranch Project Configuration

For GitHub Organization or Multibranch Pipelines, set durability per branch:

1. Open your project and click Configure.
2. Under Branch Sources, expand Property strategy.
3. Add Pipeline durability and define rules.

Example:

• main → MAX_SURVIVABILITY
• All other branches → PERFORMANCE_OPTIMIZED

3. Single Pipeline Job Configuration

In any Declarative Pipeline job:

1. Scroll to Pipeline Speed and Durability.
2. Choose the desired level:

Demonstration: Durability in Action

Create a simple pipeline that appends numbers every second. Kill Jenkins mid-run to observe each durability mode.

Pipeline Definition

pipeline {
    agent any
    stages {
        stage('For Loop - Durability Test') {
            steps {
                script {
                    sh 'touch numbers.txt'
                    for (int i = 0; i < 600; i++) {
                        sh "echo ${i} >> numbers.txt"
                        sleep 1
                    }
                }
            }
        }
    }
}

MAX_SURVIVABILITY Test

1. Set Pipeline Speed and Durability to Maximum survivability.

2. Start the build and note the output.

3. Identify and kill the Jenkins process:

   ps aux | grep -i jenkins.war
   kill -9 <PID>
   systemctl status jenkins
   

4. After restart, open the build console. The pipeline resumes at the last completed step:

   [Pipeline] sleep
   Sleeping for 1 sec
   [Pipeline] sh
   + echo 41
   

PERFORMANCE_OPTIMIZED Test

1. Switch durability to Performance optimized and rebuild.

2. Kill Jenkins mid-run as before.

3. After restart, the build fails with:

   ERROR: Cannot resume build because FlowNode 42 for FlowHead 1 could not be loaded.
   This is expected when using PERFORMANCE_OPTIMIZED and an unclean shutdown occurred.
   Finished: FAILURE
   

Note

The Jenkins dashboard will show the failed run under your project’s history.

Conclusion

• Use MAX_SURVIVABILITY for mission-critical or production pipelines requiring resumability.
• Use PERFORMANCE_OPTIMIZED when throughput is the priority and occasional rebuilds are acceptable.
• Adjust the scope—global, per-branch, or per-job—to match your workflow.

Links and References

• Jenkins Pipeline Documentation
• Managing Jenkins
• Jenkins Configuration as Code