Scheduler - PraisonAI

Schedule agents and recipes to run periodically. Schedule agents and recipes to run periodically. Perfect for monitoring tasks, periodic reports, and automated workflows.

Quick Start

Simple Usage

Scheduling a Recipe

import { recipe } from 'praisonai-ts';

// Schedule a recipe
const scheduler = await recipe.schedule('my-recipe', {
  interval: 'hourly',
  maxRetries: 3,
  timeoutSec: 300,
  maxCostUsd: 1.00,
  runImmediately: true,
});

// Start the scheduler
scheduler.start();

// Get statistics
const stats = scheduler.getStats();
console.log(`Executions: ${stats.totalExecutions}`);
console.log(`Success rate: ${stats.successRate}%`);

// Stop when done
scheduler.stop();

Scheduler Interface

interface RecipeScheduler {
  recipeName: string;
  interval: string;
  isRunning: boolean;
  
  start(): void;
  stop(): void;
  getStats(): SchedulerStats;
}

interface SchedulerStats {
  totalExecutions: number;
  successfulExecutions: number;
  failedExecutions: number;
  successRate: number;
  totalCost: number;
  lastExecutionAt?: Date;
  nextExecutionAt?: Date;
}

Using AgentScheduler Directly

import { Agent, AgentScheduler } from 'praisonai-ts';

// Create agent
const agent = new Agent({
  name: 'News Monitor',
  instructions: 'Monitor and summarize AI news',
});

// Create scheduler
const scheduler = new AgentScheduler({
  agent,
  task: 'Check for latest AI news',
  timeout: 300,
  maxCost: 1.00,
});

// Start with interval
scheduler.start({
  scheduleExpr: 'hourly',
  maxRetries: 3,
  runImmediately: true,
});

// Monitor
console.log(`Running: ${scheduler.isRunning}`);
console.log(`Stats: ${JSON.stringify(scheduler.getStats())}`);

// Stop
scheduler.stop();

Interval Formats

Format	Description
`hourly`	Every hour
`daily`	Every day
`*/30m`	Every 30 minutes
`*/6h`	Every 6 hours
`3600`	Every 3600 seconds

Configuration

TEMPLATE.yaml Runtime Block

runtime:
  schedule:
    enabled: true
    interval: "hourly"
    max_retries: 3
    run_immediately: false
    timeout_sec: 300
    max_cost_usd: 1.00

Safe Defaults

Setting	Default	Description
`interval`	hourly	Execution interval
`maxRetries`	3	Retry attempts on failure
`timeoutSec`	300	Timeout per execution
`maxCostUsd`	1.00	Budget limit

Error Handling

import { recipe, SchedulerError } from 'praisonai-ts';

try {
  const scheduler = await recipe.schedule('my-recipe', {
    interval: 'hourly',
    maxCostUsd: 0.50,
  });
  
  scheduler.start();
  
  // Monitor for errors
  scheduler.on('error', (error) => {
    console.error(`Execution failed: ${error.message}`);
  });
  
  scheduler.on('budgetExceeded', () => {
    console.log('Budget limit reached, stopping');
    scheduler.stop();
  });
  
} catch (error) {
  if (error instanceof SchedulerError) {
    console.error(`Scheduler error: ${error.message}`);
  }
}

Best Practices

Set budget limits

Prevent runaway costs

Use appropriate intervals

Match task requirements

Handle failures

Implement retry logic

Monitor statistics

Track success rates

Clean shutdown

Always call stop() when done

Scheduler CLI

CLI commands

Async Jobs

Server jobs

​Quick Start

​Scheduling a Recipe

​Scheduler Interface

​Using AgentScheduler Directly

​Interval Formats

​Configuration

​TEMPLATE.yaml Runtime Block

​Safe Defaults

​Error Handling

​Best Practices

​Related