Both catch the cron job that stopped running, and both understand a crontab line. Healthchecks.io does only that, cheaply, and you can run it yourself. okokumo watches the rest of the estate on the same plan.
Healthchecks.io is a dead man's switch. Its About page puts the goal as "do one thing and do it
well: alert when something does not happen on time", and the cron side is mature: a crontab
expression with the server's timezone, systemd OnCalendar schedules, start and
fail signals, and logs attached to pings.
An okokumo heartbeat works the same way. Your job calls a unique URL when it finishes, can call /start and /fail as well, and can attach up to 10 KB of its own output
to a failure. A cron check takes the crontab line and the IANA timezone it runs in, so the
deadline is the next real scheduled run plus your grace delay. On this axis neither product has
a feature the other lacks, and we will not pretend otherwise. Cron monitoring in detail →
Their free plan monitors twenty jobs; ours allows five checks. A hundred jobs is 20$ a month on their Business plan and 29€ on our Pro, and their Business Plus plan covers a thousand jobs for 80$ where our largest plan stops at five hundred checks for 79€. Their paid plans also include SMS, WhatsApp and phone-call credits, which we do not offer.
Healthchecks is BSD-licensed and documented for self-hosting on Python, Django and PostgreSQL or MySQL. okokumo is a hosted service only. If the monitor has to live inside your own network, or you want to read the code that decides when you are paged, that settles it in their favour.
Healthchecks.io waits for your jobs to call in and does not reach out to your services. okokumo does both on the same plan. An HTTP check makes the request itself and asserts on status, latency, body and headers, so a 200 page that says "maintenance" can still fail the check. A TLS check warns before a certificate expires and flags a broken chain. A domain check watches the registration expiry and DNS resolution. Each one counts as one check against the same plan limit as a cron job, and a failing HTTP check is confirmed from a second EU region before you are alerted.
That is the case for okokumo: the nightly backup, the API it feeds, the certificate in front of that API and the domain under it, in one place with one alerting setup. If your list is only the backup, the paragraph above about price applies instead.
okokumo runs a hosted MCP server at app.okokumo.com/mcp. Claude Code and other MCP
clients connect to it with an OAuth sign-in or an API key, then answer "is anything down?" or
"was I alerted about the backup?" from your real checks and delivery log. Healthchecks.io does not
list a first-party MCP server in its docs; several community projects wrap its management API and
run locally with a project API key. Connecting an assistant →
Both are in the EU. Their FAQ says they process and store data on Hetzner servers in Germany, and the company is registered in Latvia. Ours is in France, on Scaleway in the fr-par region, with a second EU probe region.
The difference is the contract. Their FAQ says they do not sign data processing agreements with customers, because the personal data they hold is about their customers rather than their customers' customers, which makes them a controller, not a processor, and they ask you not to put personal data in pings. That reasoning may well fit you if your jobs send nothing personal. If your procurement process requires a DPA anyway, ours is published on every plan, including Free. Security & data →
| okokumo | Healthchecks.io | |
|---|---|---|
| Free tier | 5 checks | 20 jobs |
| 100 monitored jobs | 29€/month (Pro) | 20$/month (Business) |
| Largest self-serve plan | 500 checks for 79€/month | 1,000 jobs for 80$/month |
| Crontab expression with the server's timezone | Yes | Yes, plus systemd OnCalendar schedules |
| Start and fail signals from the job | Yes, with up to 10 KB of output on a fail | Yes, with logs attached to pings |
| HTTP checks with assertions | Status, latency, body and headers | Not offered; heartbeat monitoring only, by design |
| TLS certificate and domain expiry checks | Yes, same plans | Not offered |
| SMS, WhatsApp and phone-call alerts | Not offered | Credits included on paid plans |
| Notification integrations | Email, Slack, Discord, Telegram, signed webhook | 27 listed on their homepage, PagerDuty and Opsgenie among them |
| MCP server for AI assistants | Hosted, first-party, OAuth sign-in or API key | None first-party in their docs; community servers wrap their API |
| Open source, self-hostable | Not offered | Yes, BSD 3-clause licence |
| Where monitoring data is processed | France (Scaleway fr-par) plus one EU probe region | Germany (Hetzner), per their FAQ |
| Data processing agreement | Published, every plan including Free | Not signed with customers; their FAQ says they act as controller |
Healthchecks.io plans and prices read from their pricing page on 26 October 2026, in the currency they publish and with no conversion applied. These are their monthly prices; their page offers annual billing at 20% off (16$ and 64$ a month). They also give the Business plan free to open-source projects and nonprofits. Other statements about Healthchecks.io are from their own FAQ, About page, self-hosting docs, check configuration docs and homepage integration list, read the same day. Their plans change; check before relying on this. Ours are on our own pricing page.
Twenty jobs free against our five, and a hundred jobs for 20$ a month against our 29€. If nothing on your list is an endpoint, a certificate or a domain, you would be paying us for checks you do not use. And if you want to run the monitor inside your own network, their code is BSD-licensed and documented for self-hosting; ours is not available to self-host at all.
If you monitor cron jobs and nothing else, Healthchecks.io gives you more for less. If you rely on SMS or phone calls, or on PagerDuty, Opsgenie or one of the other incident tools they integrate with, we do not have those; our channels are email, Slack, Discord, Telegram and a signed webhook. If you need to self-host, we cannot help.
Comparing per-monitor pricing as well? See okokumo vs Cronitor.