−50%

50% off your first month on any paid plan · code WELCOME50 · until Nov 30

Start free →
Guide

Cloud Cron vs AWS EventBridge: A Simpler Alternative for Scheduled HTTP Requests

You have a URL that needs to be called every few minutes. Maybe it triggers a report, clears a cache, or syncs data between two systems. On AWS, the "proper" way to do this involves EventBridge Scheduler, a Lambda function, two IAM roles, and a dead-letter queue. That's a lot of moving parts for what is essentially a scheduled curl command. And most of it exists just because EventBridge can't call an outside URL on its own.

A quick primer on EventBridge Scheduler#

Amazon EventBridge is actually two services under one name. EventBridge Rules listens for things happening inside your AWS account (a file uploaded to S3, a deployment finishing) and routes those events to other services. EventBridge Scheduler is the part that works like cron: you set a time or interval, and it triggers something.

That "something" can be almost any AWS service. Lambda functions, SQS queues, Step Functions workflows, ECS containers, about 270 targets in total. AWS added Universal Targets in 2023 to cover most of them. You can also group your schedules with Scheduler Groups for better organization.

Pricing is generous. You get 14 million free invocations per month, forever, not just the first 12 months. After that, it's $1.00 per million. For scheduling purposes, most teams never pay a cent for the Scheduler itself.

So what's the catch?

The catch: hitting an external URL is harder than it should be#

Most tutorials don't lead with this. EventBridge Scheduler can't call an outside URL directly. It's designed to talk to AWS services. If you want it to call your app's /api/cron/cleanup endpoint, you need to set up a middleman.

Option 1: go through Lambda. You write a small function that makes the HTTP request (even a three-line script that just calls fetch()). But that function needs its own security role (IAM calls it an "execution role"). Then EventBridge needs a separate security role with permission to run that function. You'll probably want a dead-letter queue too, which is a place where failed invocations get stored instead of disappearing silently. Add it all up and you're looking at 7-10 separate AWS resources just to hit one URL on a timer.

Option 2: API Destinations. This feature was built specifically for calling external URLs without Lambda. Sounds perfect, right? The problem: there's a hard 5-second timeout on the response. Not configurable, not negotiable. If your endpoint takes 6 seconds to respond, EventBridge treats it as a failure and retries. It can retry up to 185 times over 24 hours. So one slow response turns into 185 duplicate requests hammering your server.

Both options also use a non-standard cron syntax. Instead of the familiar five-field format (*/5 * * * *), EventBridge wants six fields including a year: cron(0/5 * * * ? *). You need a ? in either the day-of-week or day-of-month field. It's one of those things that wastes 20 minutes every time you set up a new schedule because you can never remember the rules.

If you're seriously considering the Lambda route, our serverless cron jobs guide covers the technical details: jitter, at-least-once delivery, cold start delays, and IAM trust policy gotchas.

What a cloud cron service does instead#

If you're looking for an AWS EventBridge alternative that skips the Lambda middleman entirely, a cloud cron service is the simple version of this. It's an external tool that calls your URLs on a schedule. No AWS account needed, no Lambda functions, no security policies to configure.

You fill in a form: the URL you want called, which HTTP method to use (GET, POST, etc.), how often to run it, and optionally a request body. Save it. Done. The service hits your endpoint on schedule, records what happened (status code, how long it took, what the server returned), and alerts you if something goes wrong.

WatchCron's Cloud Cron is one option in this space. It stores the full response body from each run, which turns out to be really useful when your endpoint returns 200 but the actual output is wrong. You can see exactly what came back without ssh-ing into your server to check logs. Alerts go through 19 different channels (email, Slack, PagerDuty, SMS, phone calls, etc.), and retries happen automatically with backoff.

Now for the obvious tradeoff: a cloud cron service only calls HTTP endpoints. It can't start an ECS container, push a message to SQS, or kick off a Step Functions workflow. Those are things EventBridge does natively. If your scheduled task lives inside AWS, EventBridge is still the shorter path. But if your task is just "call this URL," you're over-engineering it by routing through AWS.

The visibility gap nobody talks about#

This is where the difference really shows up day to day. EventBridge fires your schedule and... what happened? Did the endpoint return 200 or 500? What was in the response? EventBridge doesn't tell you any of that out of the box.

To get that information, you need to wire up CloudWatch Logs on your Lambda function, create a CloudWatch Alarm that watches for errors, connect that alarm to an SNS topic for notifications, and subscribe your email or Slack to that topic. Four extra things to set up just to answer "did my cron job work?"

With a cloud cron dashboard, every execution is a row in a list: timestamp, status code, response time, response body, error message if any. If something starts getting slower over time, response time charts make it visible before it becomes an outage. Opsgenie or Microsoft Teams get a notification within seconds of a failure, no extra plumbing required.

At 3 AM, the difference is brutal. With EventBridge, you open the AWS Console, find CloudWatch, locate the right log group, search for the failed invocation, parse the JSON output. With a cloud cron service, you open one page and scroll down.

Pricing: it's not about the invocations#

Per-invocation cost of EventBridge is essentially zero for scheduling. A job that runs every minute fires 43,200 times a month, that's a rounding error against the 14M free tier. But once you add the supporting services for external HTTP calls, costs start appearing: Lambda execution time, API Destination fees ($0.20/M), CloudWatch Logs ($0.50/GB ingested), and SNS notifications ($0.50/M). A realistic total for a few schedules with proper monitoring is $3-10/month in AWS bills, as of November 2026.

Bigger cost is time. For a team that doesn't already have EventBridge templates lying around, setting up Scheduler with proper security roles, a dead-letter queue, and CloudWatch alerting can take half a day. Even with infrastructure-as-code, someone has to write those templates and keep them updated.

A comparison for HTTP scheduling:

  • Getting started: EventBridge needs Lambda or API Destinations, security roles, and supporting resources. A cloud cron service needs one form.
  • How often: both go down to 1-minute intervals.
  • Timeout: API Destinations max out at 5 seconds (hard limit). WatchCron Cloud Cron gives you 60.
  • Seeing what happened: EventBridge doesn't store HTTP responses. Cloud Cron keeps the full body.
  • Getting alerted: EventBridge requires SNS + CloudWatch setup. Cloud Cron has 19 channels built in.
  • Cron format: EventBridge uses a 6-field format with a mandatory year. Everyone else uses 5 fields.

When EventBridge is the right choice#

If your scheduled task talks to an AWS service, not an HTTP endpoint, EventBridge is the obvious tool. Starting a Step Functions workflow that orchestrates a multi-step data pipeline, pushing messages into SQS for background workers, spinning up an ECS container on a timer: EventBridge handles all of that without any intermediary. One schedule, one target, done.

It's also the right pick when compliance rules require everything to stay inside AWS. Some organizations won't let you route scheduled tasks through external services, and that's a legitimate constraint. And if you have thousands of one-time "do this at exactly 3 PM tomorrow" tasks, the at() expression was designed for exactly that pattern.

Where it stops making sense: you're running a WordPress site, a Laravel app, a SaaS product on a VPS, or anything where the scheduled task is just "hit this URL." For that, EventBridge adds infrastructure you'll spend more time maintaining than the task itself takes to run.

Moving your HTTP schedules off EventBridge#

If you've already got EventBridge calling URLs through Lambda, switching to an EventBridge alternative doesn't have to be scary. For each schedule:

  1. Write down the URL, HTTP method, any headers, and the cron expression (convert it from 6-field to 5-field).
  2. Create a task in WatchCron or any cloud cron service with those same settings.
  3. Run both in parallel for a day or two. Compare the execution logs to make sure they match.
  4. Once it checks out, turn off the EventBridge schedule. Give it a week, then delete the Lambda function, the security roles, and the dead-letter queue.

If you're already using WatchCron for uptime monitoring, SSL checks, or heartbeat monitoring, the scheduled tasks just become another row in the same dashboard. For the GCP version of this comparison, we covered Google Cloud Scheduler separately.

Plenty of teams end up running a hybrid setup: EventBridge handles the AWS-native stuff (Step Functions triggers, SQS pushes), and a cloud cron service handles all the HTTP endpoints. Each tool does what it was built for, and for the HTTP side, a cloud cron service is a cleaner AWS EventBridge alternative than wrapping everything in Lambda. And you finally get to delete those functions that were really just glorified curl wrappers.

Frequently asked#

They're part of the same service but do different things. EventBridge Rules reacts to events (something happened, now do something). EventBridge Scheduler runs on a clock (do this every 5 minutes, or at 3 PM tomorrow). For cron-like scheduling, Scheduler is the one you want.

Not directly. You either write a Lambda function that makes the HTTP call (which means setting up the function, two security roles, and ideally a dead-letter queue), or use API Destinations (which have a hard 5-second timeout on the response that can't be changed).

The scheduler invocations are almost free: 14 million per month at no charge, $1 per million after that. The real cost comes from the Lambda functions, CloudWatch logging, and SNS notifications you need around it. A realistic total for a few schedules with monitoring is $3-10/month.

For HTTP-based scheduled tasks, yes. You give it a URL and a schedule, and it handles execution, logging, retries, and alerting. For AWS-native targets like SQS queues or Step Functions, EventBridge is still the better tool since it talks to those services directly.

It can't call external URLs without a middleman. The Lambda route adds complexity. The API Destination route has a 5-second timeout you can't change. Either way, you're setting up extra infrastructure for something that should be simple.

Yes. Cloud cron services like WatchCron let you schedule HTTP requests without any AWS infrastructure. Enter a URL, pick a schedule, and the service handles execution, logging, retries, and alerts through 19 notification channels.

Written by Vitalii Holben

Founder of WatchCron. Full-stack developer building monitoring tools that help teams keep cron jobs, uptime, and SSL certificates under control. Based in Europe, shipping code since 2014.

Start monitoring in under 2 minutes.

Free plan includes 20 checks · No credit card required
Pricing Docs
Log in Start free →