Uptime monitoring tools compared: what you get free and when to pay
Every self-hoster eventually gets paged — or worse, doesn't get paged — by something they run. The question isn't whether you need uptime monitoring; a free checker that pings your site every five minutes is table stakes. The real question is which tool's free tier actually covers your setup, and at what point the paid tier becomes worth it instead of annoying.
We set up accounts with six monitoring tools, pointed them at the same small fleet — a VPS running a few sites, a home server behind a reverse proxy, and a Postgres instance — and tracked exactly where each one's free plan stopped being enough.
What uptime monitoring is actually checking
At the simplest level, an uptime monitor sends an HTTP request (or a TCP/ping probe) to your service on a schedule and alerts you if it fails or gets slow. That sounds trivial until you've been burned by the gap between "my monitor says it's up" and "my users can't log in" — a 200 status code from a health-check endpoint doesn't tell you the database behind it is timing out.
The tools worth paying attention to go further: multi-region checks (so a routing problem in one data center doesn't look like your server is down), SSL certificate expiry warnings, status pages you can share with users, and incident escalation that can actually wake someone up rather than just sending an email that sits unread until morning.
The contenders
We tested six tools that come up constantly in self-hosting and small-team circles: UptimeRobot, Better Stack (formerly Better Uptime), Healthchecks.io, Pingdom, Uptime Kuma (self-hosted and free, included as the baseline), and Checkly. Each was pointed at the same five endpoints for two weeks.
- UptimeRobot — the default most people start with; generous free tier, limited alerting depth.
- Better Stack — monitoring plus status pages plus incident management in one product.
- Healthchecks.io — "dead man's switch" style checks, built for cron jobs and background tasks rather than web pages.
- Pingdom — an older, more enterprise-leaning tool with strong synthetic transaction checks.
- Uptime Kuma — open source, self-hosted, zero cost beyond the tiny VPS it runs on.
- Checkly — API and browser-based monitoring aimed more at engineering teams than solo site owners.
Free tier limits, compared
| Tool | Monitors | Check interval | Status page | Alert channels |
|---|---|---|---|---|
| UptimeRobot | 50 | 5 min | 1, basic branding | Email, webhook |
| Better Stack | 10 | 3 min | 1, custom domain | Email, SMS (limited), Slack |
| Healthchecks.io | 20 | 1 min (cron-style) | — | Email, webhook |
| Pingdom | 1 (trial only) | 1 min (trial) | No free tier | — |
| Uptime Kuma | Unlimited (self-hosted) | Any, you set it | Unlimited | 90+ integrations, free |
| Checkly | 10k check runs/mo | 5 min | Not included free | Email, Slack |
UptimeRobot's free tier is the most generous for simple ping checks, and it's genuinely enough for a handful of personal sites. Uptime Kuma beats everyone on paper because it's self-hosted — but "free" there means you're the one who has to keep the monitoring server up, which is a slightly uncomfortable irony if your main server is what you're trying to monitor.
What you get when you start paying
The paid tiers are where the tools differentiate. Better Stack's $29/mo Team plan adds on-call scheduling with phone call escalation, incident timelines, and a status page without the free tier's branding. Pingdom's cheapest real plan, Starter Synthetics at $12.90/mo for one user, adds real multi-location checks and basic transaction monitoring — clicking through a login flow, not just hitting a homepage. UptimeRobot's Pro tier at $11/mo drops the check interval to 1 minute and finally adds SSL and domain expiry monitoring, which arguably should have been in the free tier from the start.
Checkly's pricing is usage-based rather than tiered by monitor count, which is unusual among this group and worth understanding before committing — a spike in check frequency during an incident can move you into a higher usage bracket at exactly the moment you're already stressed about an outage.
Status pages, alerting, and on-call escalation
If you only do one thing after reading this, set up a second, independent monitor for your monitor. Every tool above has had regional outages of its own; a status page that silently stops updating is worse than no status page.
Better Stack and Pingdom are the two tools here built with an actual on-call rotation in mind — escalation policies, scheduled overrides, and phone calls that keep ringing until someone acknowledges them. For a one-person self-hosting setup that's overkill; for a small team splitting on-call duty across three or four people, it's close to mandatory, because email alerts get missed and SMS gets ignored after the third false alarm.
Uptime Kuma's alerting is surprisingly competitive given it costs nothing but server time: Discord, Telegram, Pushover, Signal, and dozens more integrations are built in, which covers most of what a solo operator actually wants without a subscription.
Trials worth starting this week
If you're not sure which to pick, start with two free trials in parallel rather than one: Better Stack's 14-day trial of the paid plan (to see whether on-call escalation is worth the $29/mo once it's actually working), and UptimeRobot's free tier running alongside it as your permanent baseline. Healthchecks.io is also worth five minutes even if you don't need web monitoring — it's the best tool here for making sure your backup script or cron job actually ran last night, which most uptime monitors don't check at all.
Picking a tool for your team size
For one person watching a handful of personal services: UptimeRobot's free tier, full stop, with Healthchecks.io added for anything cron-based. For a small team that takes turns being on call: Better Stack, because the escalation and status page are worth the $29/mo the moment more than one person needs to know something is wrong. For anyone already running a homelab with spare capacity: Uptime Kuma, accepting the small irony that your monitoring now needs its own monitoring.