> ## Documentation Index
> Fetch the complete documentation index at: https://docs.vertracloud.app/llms.txt
> Use this file to discover all available pages before exploring further.

# Run Scheduled Jobs

> Run recurring tasks, like cleanups, reports and reminders, from an always-on background application with an in-process scheduler.

Vertra Cloud keeps your process running around the clock, so a recurring task is simply a scheduler
inside your code. A job that doesn't serve HTTP runs as a background application, which doesn't need
web publishing and fits the smaller plans.

## Choose where the job lives

* **Inside your existing app**: simplest, if the job is light and shares code with it.
* **In a separate background application**: better if the job is heavy or you don't want it to
  compete with requests for memory. It restarts and scales on its own.

## Node.js

```bash theme={null}
npm install node-cron
```

```javascript index.js theme={null}
const cron = require('node-cron');

cron.schedule('0 3 * * *', async () => {
  console.log('nightly cleanup started');
  await cleanup();
}, { timezone: 'America/Sao_Paulo' });

console.log('scheduler running');
```

## Python

```bash theme={null}
pip install apscheduler
```

```python main.py theme={null}
from apscheduler.schedulers.blocking import BlockingScheduler

scheduler = BlockingScheduler(timezone="America/Sao_Paulo")

@scheduler.scheduled_job("cron", hour=3)
def nightly_cleanup():
    print("nightly cleanup started", flush=True)

scheduler.start()
```

`BlockingScheduler` keeps the process alive. With a script that ends after scheduling, the process
exits with code `0` and isn't restarted.

## Things to get right

<AccordionGroup>
  <Accordion title="Always set the timezone">
    Don't depend on the container's clock zone: pass the timezone to the scheduler explicitly, as in
    the examples.
  </Accordion>

  <Accordion title="Make jobs safe to run twice">
    A restart during a job means it may run again. Record what was processed, or design the job so a
    second run changes nothing.
  </Accordion>

  <Accordion title="Catch errors inside the job">
    An unhandled error can take the whole process down, and 5 crashes in 10 minutes pause
    auto-restart for 24 hours. Wrap the job body in `try`/`catch` and log the failure.
  </Accordion>

  <Accordion title="Log every run">
    Print a line when a job starts and ends; the **Logs** tab is where you'll check that it ran. In Python
    use `flush=True` so the line appears immediately.
  </Accordion>
</AccordionGroup>