Questions
What this costs, how a failure is confirmed before anybody is contacted, and what happens when one of our own machines is the thing that broke. Every check has its own page with the questions particular to it.
Plans and price
Is there a free plan, and what is in it?
Yes, and it is a plan rather than a trial: you can stay on it for as long as you like. It holds 5 monitors, checked as often as every 5 min, with 1 seat and 90 days of history. It checks websites and APIs, DNS, TCP ports and ping. Mail servers, SSL/TLS certificates, blocklists, heartbeats and WebSockets are on the paid plans, because they cost more to run: a single email blocklist check is fifty lookups.
Do I need a card to start?
No. The free plan asks for an email address and a password, and nothing else. You are asked for a card when you choose a paid plan, and not before.
What do the paid plans cost?
Starter is €10 a month, Pro is €32 and Business is €77. Paid yearly they are €9, €29 and €69 a month. They differ by how many monitors you may have (30, 100 and 300), how often they may be checked, how many people can sign in, and which alert channels are open. Every plan gets every check from three continents; none of that is held back for the expensive ones.
How long do you keep the history?
Raw check results for 90 days on every plan, rolled up into hourly summaries. The summaries, the incidents and the record of what was sent are kept for 90 days on Free and Starter, and for a year on Pro and Business.
How the checking works
How do you avoid false alarms?
One probe failing tells you about one probe, and most single failures are the network in between rather than your site. So a failure is not an outage until a second probe agrees, and that second probe has to be on a different continent and at a different hosting provider. A failure nobody confirms is recorded, and visible, and never alerted on.
What happens if one of your own probes goes down?
Silence from a probe is recorded as no data, never as an outage. This is the failure that matters most to get right: if a dead machine of ours counted as a failure, one VPS losing its network would page every customer in that region at once, with total confidence, about sites that are perfectly fine.
Where do you check from?
Three continents: Amsterdam (Europe), New York (North America), Sydney (Oceania). Two hosting providers, so a confirmation is never two machines sharing one company's bad day. Every plan is checked from all of them.
How often do you check?
As often as every 30 sec on Pro and Business, every 1 min on Starter and every 5 min on Free. You choose per monitor, and slower is often the right answer: a certificate is a date, so we read it twice a day rather than every minute.
What can you monitor?
Ten kinds of check: https, smtp, dns, email blocklists, ssl/tls, ping, tcp port, heartbeat, domain blocklists, websocket. Each one has its own page explaining what it catches and what it cannot.
Alerts
How do I get told?
Email and webhooks on every plan, including the free one. PagerDuty from Starter. Chat channels and an on-call rota from Pro, because those are a team's tools and Pro is where teams start.
Will I be woken up for planned work?
Not if you say so first. A maintenance window covers the monitors or groups you choose, for the time you choose, and it can repeat. During one the check still runs and the result is still recorded; it is left out of the uptime figure and nobody is alerted, because time you chose is not time you suffered.
What if one outage takes twenty things down with it?
You are told about the cause, once. A monitor can depend on another one, and when the thing above it is already known to be down the ones below it still open incidents and still appear in the history, but nobody is paged for them. The number of alerts you did not get is shown beside the one you did.
The rest of it
Can I put a status page in front of my customers?
Yes, from Starter. It publishes the monitors you pick, with their history, at an address you can send to somebody. Pro lets you choose that address and Business allows 10 pages, which is what an agency with ten clients needs.
Is there an API?
Yes, on every plan, and it is the same API the interface uses. Bearer tokens you create yourself, an OpenAPI document, a command line client that is one Python file with no dependencies, and an MCP server so an assistant can read your monitors and incidents directly.
How is this different from other uptime monitors?
Confirmation from a second continent before anybody is contacted, rather than after somebody complains about the alert. Most monitoring sends what one machine saw and leaves you to work out whether it was real, and the cost of that is not the false alarm itself, it is that you stop believing the true ones.
The checks
What each one catches, what it cannot, and the questions particular to it.