Cloudflare Cron Triggers

Status: Production Ready ✅ Last Updated: 2025-11-25 Dependencies: cloudflare-worker-base (for Worker setup) Latest Versions: wrangler@4.50.0, @cloudflare/workers-types@4.20251125.0

Quick Start (5 Minutes)

1. Add Scheduled Handler to Your Worker

src/index.ts:

export default {
  async scheduled(
    controller: ScheduledController,
    env: Env,
    ctx: ExecutionContext
  ): Promise<void> {
    console.log('Cron job executed at:', new Date(controller.scheduledTime));
    console.log('Triggered by cron:', controller.cron);

    // Your scheduled task logic here
    await doPeriodicTask(env);
  },
};

Why this matters:

• Handler must be named exactly scheduled (not scheduledHandler or onScheduled)
• Must be exported in default export object
• Must use ES modules format (not Service Worker format)

2. Configure Cron Trigger in Wrangler

wrangler.jsonc:

{
  "name": "my-scheduled-worker",
  "main": "src/index.ts",
  "compatibility_date": "2025-10-23",
  "triggers": {
    "crons": [
      "0 * * * *"  // Every hour at minute 0
    ]
  }
}

CRITICAL:

• Cron expressions use 5 fields: minute hour day-of-month month day-of-week
• All times are UTC only (no timezone conversion)
• Changes take up to 15 minutes to propagate globally

3. Test Locally

# Enable scheduled testing
bunx wrangler dev --test-scheduled

# In another terminal, trigger the scheduled handler
curl "http://localhost:8787/__scheduled?cron=0+*+*+*+*"

# View output in wrangler dev terminal

Testing tips:

• /__scheduled endpoint is only available with --test-scheduled flag
• Can pass any cron expression in query parameter
• Python Workers use /cdn-cgi/handler/scheduled instead

4. Deploy

npm run deploy
# or
bunx wrangler deploy

After deployment:

• Changes may take up to 15 minutes to propagate
• Check dashboard: Workers & Pages > [Your Worker] > Cron Triggers
• View past executions in Logs tab

When to Load References

Load immediately when user mentions:

• cron-expressions-reference.md → "cron syntax", "schedule format", "expression", "minute hour day", "every X minutes"
• common-patterns.md → "examples", "use cases", "patterns", "real-world", "database cleanup", "report generation", "how to"
• integration-patterns.md → "implement", "Hono", "multiple triggers", "bindings", "workflows", "error handling"
• wrangler-config.md → "configuration", "wrangler.jsonc", "multiple crons", "environment-specific", "dev staging production"
• testing-guide.md → "test", "local development", "__scheduled", "unit test", "curl", "debugging"

Load proactively when:

• Building new scheduled task → Load integration-patterns.md
• Configuring wrangler.jsonc → Load wrangler-config.md
• Debugging cron expression → Load cron-expressions-reference.md
• Testing locally → Load testing-guide.md
• Looking for examples → Load common-patterns.md

Cron Expression Syntax

Five-Field Format

* * * * *
│ │ │ │ │
│ │ │ │ └─── Day of Week (0-6, Sunday=0)
│ │ │ └───── Month (1-12)
│ │ └─────── Day of Month (1-31)
│ └───────── Hour (0-23)
└─────────── Minute (0-59)

Special Characters

Common Patterns

# Every minute
* * * * *

# Every 5 minutes
*/5 * * * *

# Every 15 minutes
*/15 * * * *

# Every hour at minute 0
0 * * * *

# Every hour at minute 30
30 * * * *

# Every 6 hours
0 */6 * * *

# Every day at midnight (00:00 UTC)
0 0 * * *

# Every day at noon (12:00 UTC)
0 12 * * *

# Every day at 3:30am UTC
30 3 * * *

# Every Monday at 9am UTC
0 9 * * 1

# Every weekday at 9am UTC
0 9 * * 1-5

# Every Sunday at midnight UTC
0 0 * * 0

# First day of every month at midnight UTC
0 0 1 * *

# Twice a day (6am and 6pm UTC)
0 6,18 * * *

# Every 30 minutes during business hours (9am-5pm UTC, weekdays)
*/30 9-17 * * 1-5

CRITICAL: UTC Timezone Only

• All cron triggers execute on UTC time
• No timezone conversion available
• Convert your local time to UTC manually
• Example: 9am PST = 5pm UTC (next day during DST)

ScheduledController Interface

interface ScheduledController {
  readonly cron: string;           // The cron expression that triggered this execution
  readonly type: string;           // Always "scheduled"
  readonly scheduledTime: number;  // Unix timestamp (ms) when scheduled
}

Properties

controller.cron (string)

The cron expression that triggered this execution.

export default {
  async scheduled(controller: ScheduledController, env: Env): Promise<void> {
    console.log(`Triggered by: ${controller.cron}`);
    // Output: "Triggered by: 0 * * * *"
  },
};

Use case: Differentiate between multiple cron schedules (see Multiple Cron Triggers pattern).

controller.type (string)

Always returns "scheduled" for cron-triggered executions.

if (controller.type === 'scheduled') {
  // This is a cron-triggered execution
}

controller.scheduledTime (number)

Unix timestamp (milliseconds since epoch) when this execution was scheduled to run.

export default {
  async scheduled(controller: ScheduledController): Promise<void> {
    const scheduledDate = new Date(controller.scheduledTime);
    console.log(`Scheduled for: ${scheduledDate.toISOString()}`);
    // Output: "Scheduled for: 2025-10-23T15:00:00.000Z"
  },
};

Note: This is the scheduled time, not the actual execution time. Due to system load, actual execution may be slightly delayed (usually <1 second).

Execution Context

export default {
  async scheduled(
    controller: ScheduledController,
    env: Env,
    ctx: ExecutionContext  // ← Execution context
  ): Promise<void> {
    // Use ctx.waitUntil() for async operations that should complete
    ctx.waitUntil(logToAnalytics(env));
  },
};

ctx.waitUntil(promise: Promise<any>)

Extends the execution context to wait for async operations to complete after the handler returns.

Use cases:

• Logging to external services
• Analytics tracking
• Cleanup operations
• Non-critical background tasks

export default {
  async scheduled(controller: ScheduledController, env: Env, ctx: ExecutionContext): Promise<void> {
    // Critical task - must complete before handler exits
    await processData(env);

    // Non-critical tasks - can complete in background
    ctx.waitUntil(sendMetrics(env));
    ctx.waitUntil(cleanupOldData(env));
    ctx.waitUntil(notifySlack({ message: 'Cron completed' }));
  },
};

Important: First waitUntil() that fails will be reported as the status in dashboard logs.

Integration Patterns

6 production-ready cron patterns:

1. Standalone Worker with Cron - Single scheduled function for background tasks (database cleanup, report generation)
2. Hono + Cron Combination - HTTP endpoints + scheduled tasks in one Worker, sharing bindings and reducing costs
3. Multiple Cron Triggers - Different schedules for different tasks using controller.cron to route execution
4. Accessing Bindings - Use D1, KV, R2, AI, Vectorize, Queues, Workflows, Durable Objects in scheduled functions
5. Integrating with Workflows - Trigger complex, long-running multi-step workflows on schedule
6. Error Handling Best Practices - Comprehensive error handling with retry logic, alerting (Slack/email), failure logging, and monitoring

Load references/integration-patterns.md for complete implementations with code examples, configuration details, and best practices.

Wrangler Configu