Monitoring Vercel Cron: know whether the job ran

Setup

Set up in three steps

1. Create a heartbeat

   In your account, create a new one under “Heartbeats”, with the rhythm from your vercel.json. On the Hobby plan give it generous leeway, at least two hours, because Vercel starts the job some time within the hour. You receive a ping address of its own, such as https://runcheck.de/a8k3m2x9q1.

2. The check-in at the end of the route

   Open the route listed under crons in vercel.json and put the check-in after the actual work, before the response. The CRON_SECRET check comes from the Vercel documentation and stays as it is.

   // app/api/cron/route.ts
   import type { NextRequest } from 'next/server';
   
   export async function GET(request: NextRequest) {
     const authHeader = request.headers.get('authorization');
     if (!process.env.CRON_SECRET || authHeader !== `Bearer ${process.env.CRON_SECRET}`) {
       return new Response('Unauthorized', { status: 401 });
     }
   
     // ... your work ...
   
     await fetch('https://runcheck.de/a8k3m2x9q1', { signal: AbortSignal.timeout(10000) }).catch(() => {});
     return Response.json({ success: true });
   }

3. Deploy and trigger once

   Vercel only invokes cron jobs for the production deployment. After deploying, trigger the job by hand in the dashboard from the deployment summary, or on the command line with vercel crons run /api/cron (currently beta, according to Vercel). In your Cronjob.de account, “Waiting for the first check-in …” turns green.

The ready-made lines for Linux, PHP, Python, Node.js, n8n, Make, Supabase and Vercel are on the page Cron job monitoring. Your own ping address appears in your account once you create the heartbeat, with the same lines ready to copy.

When something fails

What happens when a run fails

If the invocation does not arrive, the work throws an exception or the function runs into its time limit, the last line is never reached: the check-in is missing. After the leeway we alert you; the next successful run brings the all-clear.

FAQ

Common questions about Vercel Cron

What does a cron job entry in vercel.json look like?

Like this, with your route's path and the schedule in UTC: { "crons": [ { "path": "/api/cron", "schedule": "0 5 * * *" } ] }

How often can a cron job run on Vercel?

According to Vercel, at most once a day on the Hobby plan; more frequent expressions fail during deployment. Pro and Enterprise allow up to once per minute (as of October 2026). If a route should run more often without changing your plan, Cronjob.de can call it directly: enter the address, add the header Authorization: Bearer with your CRON_SECRET, choose the schedule. For free that works in 5-minute steps, from Plus every minute. Each call then counts at Vercel like an ordinary function invocation.

Why is there an await before the fetch?

So the check-in goes out before the function responds. Then it really means: the work is done. AbortSignal.timeout limits the wait to ten seconds, and an error while checking in is ignored and does not affect your job.

Does this work with the Pages Router or another framework?

Yes. The check-in is an ordinary fetch call; fetch and AbortSignal.timeout are built in since Node.js 18. What matters is the place: after the work, before the response.