node-cron Monitoring — Catch Silent Failures in Node.js Scheduled Jobs | Pakyas

The problem

node-cron schedules jobs inside your Node.js process, which is exactly why its failures are so easy to miss. If the process crashes, gets OOM-killed, is redeployed, or never restarts, cron.schedule() simply never fires — and there is no log line, no exception, and no signal that anything is wrong. Even when the process is alive, an unhandled rejection inside the task callback can swallow the error so the job appears to have run when it never did. A try/catch that logs to a file nobody watches is not monitoring. The result is the classic silent failure: a nightly backup or sync that quietly stops running for days, with a green-looking app the whole time. Silence is not success, and a job that never fired is Missing — not failed — but without an external observer you cannot tell the difference.

How Pakyas helps

Pakyas uses execution-signal monitoring: instead of polling your server, it waits for your node-cron task to prove it ran by sending a signal at the start and end of each run. Pakyas compares those signals against the job's expected schedule and tells you precisely what happened — On Schedule, Late (the signal arrived but later than expected), Missing (the signal never arrived, so the process likely died or was redeployed), Overrunning (the job started but never reported completion in time), or Error (the job ran and explicitly reported a failure). Because the signal comes from inside your task callback, a crashed process or a redeploy that stops the scheduler shows up as Missing — the exact case a log line can never catch. You get one clear alert that means action, not a dashboard you have to remember to check.

Set it up

1. 1

   Create a check and copy its ping URL

   # In Pakyas, create a check for your scheduled job and copy its ping URL.
   # It has the form (public_id is the check's UUID):
   #   https://ping.pakyas.com/{public_id}
   # Start signal:   https://ping.pakyas.com/{public_id}/start
   # Failure signal: https://ping.pakyas.com/{public_id}/fail

   Set the expected schedule on the check to match your node-cron cron expression (e.g. daily at 02:00). Pakyas uses it to detect Late and Missing runs.

2. 2

   Add the ping URL to your environment

   # .env
   PAKYAS_PING_URL=https://ping.pakyas.com/your-check-public-id

   Keep the public_id out of source control. Node 18+ has fetch built in, so no extra dependency is needed.

3. 3

   Signal start, success, and failure from inside the task

   const cron = require('node-cron');
   
   const PING = process.env.PAKYAS_PING_URL;
   
   // Daily at 02:00 server time
   cron.schedule('0 2 * * *', async () => {
     // Signal that the job started (enables Overrunning detection)
     await fetch(`${PING}/start`).catch(() => {});
   
     try {
       await runBackup(); // your actual job
       // Success signal (default endpoint, no modifier)
       await fetch(PING).catch(() => {});
     } catch (err) {
       // Explicit failure signal
       await fetch(`${PING}/fail`).catch(() => {});
       throw err;
     }
   });

   The .catch(() => {}) keeps a transient ping network error from crashing your job. The cron expression here must match the schedule you set on the Pakyas check.

Pricing

Pakyas has four tiers: Free ($0), Developer ($9/mo), Pro ($29/mo), and Business ($99/mo). The Free tier monitors up to 10 checks, which is enough to cover the scheduled jobs in a typical Node.js app at no cost.

See the full breakdown on the pricing page.