How to Monitor Your App Uptime and Performance Without Paying a Cent

Your app goes down at 3am. Do you find out from a monitoring alert, or from a user tweeting at you? For a solo builder, the second one is the default. Free monitoring fixes that, and you need exactly three kinds of it: something that checks the site is up, something that catches background jobs that silently stop, and something that tells you when your code throws errors.

Here is a free option for each, with the limits that matter.

1. Is the site up?

UptimeRobot

The free plan gives you 50 monitors checked every 5 minutes, one basic status page, 3 months of data retention and 5 integrations. Monitor types include HTTP, port, ping, keyword, API, DNS and more. No card is required. The catch is the license: the free plan is for hobby and nonprofit projects, and commercial use is not allowed. A Hacker News thread described this as a terms change after people had signed up. If your project earns money, plan to pay or choose another tool. See UptimeRobot or visit uptimerobot.com.

Better Stack

10 monitors and heartbeats, 1 status page, and Slack and email alerts. It also bundles logs, traces and web events at 3 GB a month each, kept for 3 days. The plan is labeled free for personal projects, and paid plans start at $29 a month, which one G2 reviewer called steep for small open source services. See Better Stack or visit betterstack.com.

Instatus

A public status page, 15 monitors checked every 2 minutes, email alerts, 5 team members and 200 subscribers, with flat pricing and no per seat fees. Alerts are email only, and SMS needs Pro. If you want a status page your users can bookmark, this is the nicest free one here. See Instatus or visit instatus.com.

2. Did my background job run?

Uptime checks cannot see a nightly backup that quietly stopped. These tools work the other way around: your job pings them when it finishes, and they alert you if the ping does not arrive.

Healthchecks.io

The Hobbyist plan includes 20 jobs and 100 log entries per job. You ping a unique URL with a plain curl request, with start, finish and failure signals. Busy jobs lose log history quickly, and SMS and phone calls start on the $20 Business plan. It is open source and can be self hosted. See Healthchecks.io or visit healthchecks.io.

Cronitor

The Hacker plan is free forever with 5 monitors for jobs, checks and heartbeats, 5 minute checks, 1 user and one basic status page with 50 subscribers. Alerts go to email, Slack, Discord, Teams and webhooks. SMS, on call integrations and branded status pages are paid. See Cronitor or visit cronitor.io.

3. What broke in my code?

Sentry

The Developer plan is $0 for one user and unlimited projects, with 5,000 errors a month, 5 GB of logs, 5 million tracing spans, 50 session replays, 1 uptime monitor and 30 days of lookback. Unlimited users start on Team at $26 a month billed annually. Pricing is the top complaint on G2, and one reviewer said the error budget runs out often, which stops monitoring until it resets. A noisy bug can eat 5,000 errors in an afternoon, so set a sampling rate. See Sentry or visit sentry.io.

Bugsnag

The free plan is for one solo user, with 7,500 events a month, 1 million performance spans and 7 days of retention. That retention is short for a bug that shows up weekly. Paid plans keep 60 days. See Bugsnag or visit bugsnag.com.

A zero dollar monitoring setup

• Site checks: Better Stack or Instatus for a hobby project, a paid plan once it earns money.
• Scheduled jobs: Healthchecks.io, which has the most room.
• Errors: Sentry, with a sample rate set so one bug does not use your quota.
• Status page: Instatus.

Then test it. Take your site down on purpose for six minutes, and confirm that an alert reaches your phone. A monitor you have never seen fire is a guess.