PipelinePulse

A lightweight yet powerful CI/CD orchestration tool with progressive rollouts, metric-based deployments, and self-healing capabilities.

Why PipelinePulse?

PipelinePulse integrates canary deployments and auto-rollbacks into your existing CI/CD pipeline without requiring a complete overhaul.

System Architecture

graph TD
    CI[Your CI System] -->|Triggers| PP[PipelinePulse]
    PP -->|Build/Test Commands| CI
    PP -->|Deploy Commands| INFRA[Infrastructure<br/>K8s/Cloud/etc]
    PP -->|Records State| DB[(Deployment<br/>History DB)]
    PP -->|Reads Metrics| METRICS[Metrics Source]
    METRICS -->|Feedback| PP
    PP -->|Auto-Rollback| INFRA

Getting Started

• Install the package: npm install pipelinepulse
• Create a configuration file (see Quick Start below)
• Run your pipeline with npx pipelinepulse start pipeline.yaml
• For programmatic usage, see the examples below

Features

Core Orchestration with Resilience

• Exponential backoff retry logic for command execution
• Concurrent execution of pipeline stages with configurable parallelism
• Timeout capability to prevent hung deployments
• Recovery from network failures and transient errors

Metric-Driven Deployments

• Progressive traffic percentage rollouts
• Metric evaluation with automated rollback thresholds
• Health checks during deployment

Self-Healing Auto-Rollback

• Automatic rollback when metrics exceed thresholds
• Configurable error thresholds with baseline comparison
• Immediate response to quality degradation without human intervention

State Persistence

• SQLite-based state persistence for deployment history
• Automatic recovery from crashes
• Comprehensive tracking of pipeline and stage execution

Observability

• Structured logging with component tagging
• Metrics for deployment success/failure rates and durations
• Lead time tracking for deployments

Security

• Injection Prevention: Validates commands to prevent security risks
• Config Validation: Blocks patterns like curl | bash via pattern matching
• YAML Validation: Checks against MITRE ATT&CK patterns

Installation

npm install pipelinepulse

Quick Start

Create a configuration file pipeline.yaml:

pipeline:
  name: simple-deployment
  maxConcurrency: 2
  maxRetries: 3
  stages:
    - name: build
      command: npm run build
    - name: test
      command: npm test
    - name: deploy
      command: npm run deploy

canary:
  enabled: true
  initialPercentage: 20
  samplingPeriod: 30
  errorThresholdPercentage: 5
  metrics:
    - error-rate
    - response-time
  metricsCommands:
    error-rate: node scripts/get-error-rate.js
    response-time: node scripts/get-response-time.js
  deployCommands:
    initial: node scripts/deploy-initial.js --percentage=20
    update: node scripts/deploy-update.js
    final: node scripts/deploy-final.js
  rollbackCommands:
    execute: node scripts/rollback.js

Run your pipeline:

npx pipelinepulse start pipeline.yaml

Usage Examples

Programmatic Usage

const PipelinePulse = require('pipelinepulse');

async function deployApp() {
  const pulse = new PipelinePulse({
    configPath: './pipeline.yaml',
    dbPath: './deployment-history.db'
  });
  
  await pulse.initialize();
  
  const pipeline = await pulse.createPipeline({
    name: 'production-deploy',
    maxConcurrency: 2
  });
  
  pipeline.addStage({ name: 'build', command: 'npm run build' });
  pipeline.addStage({ name: 'test', command: 'npm test' });
  pipeline.addStage({ name: 'deploy', command: 'npm run deploy' });
  
  const result = await pipeline.execute();
  
  if (result.success) {
    console.log(`Deployment completed in ${result.duration / 1000} seconds`);
  } else {
    console.error(`Deployment failed at stage: ${result.failedStage}`);
  }
  
  await pulse.stop();
}

deployApp().catch(console.error);

Canary Deployment Example

const canary = await pulse.createCanaryDeployment({
  initialPercentage: 10,
  samplingPeriod: 30,
  errorThreshold: 5
});

const result = await canary.deploy({
  service: 'payment-service',
  newVersion: '2.1.0',
  deployCommands: {
    initial: 'node scripts/deploy.js --version=2.1.0 --percentage=10',
    update: 'node scripts/increase-traffic.js --percentage=50',
    final: 'node scripts/increase-traffic.js --percentage=100'
  },
  rollbackCommands: {
    execute: 'node scripts/rollback.js'
  },
  metricsCommands: {
    'error-rate': 'node scripts/get-error-rate.js',
    'response-time': 'node scripts/get-response-time.js'
  }
});

console.log(`Canary deployment ${result.success ? 'succeeded' : 'failed'}`);

CLI Commands

pipelinepulse <command> [options]

Commands:

• start <config-path> → Start the PipelinePulse server
• deploy <config-path> <name> → Execute a pipeline from config
• status <pipeline-id> [port] → Check pipeline status
• help → Show help

API Reference

PipelinePulse Class

const pulse = new PipelinePulse(options);
await pulse.initialize();
const pipeline = await pulse.createPipeline(config);
const canary = await pulse.createCanaryDeployment(config);
await pulse.stop();

Pipeline Class

pipeline.addStage({ name, command, args, timeout });
const result = await pipeline.execute();

CanaryDeployment Class

const result = await canary.deploy({
  pipelineId,
  service,
  newVersion,
  deployCommands,
  rollbackCommands,
  metricsCommands
});

Chaos Engineering & Resilience Testing

PipelinePulse is built with resilience in mind and has been tested against various failure scenarios:

• Network Outages: Recovers automatically from temporary network failures during metric collection or command execution
• Process Termination: Maintains state even when abruptly terminated, allowing recovery on restart
• Resource Contention: Handles high concurrency with configurable limits to prevent resource exhaustion
• Metric Anomalies: Properly distinguishes between normal variation and severe issues