What it does.
Nine kinds of monitor, nine ways to be told, checked from three locations in the EU. Every plan gets all of it.
Every kind on every plan
Monitor kinds.
Check every 30 s at the fastest, or once an hour. The fastest interval depends on the plan.
HTTP
Requests a URL, follows up to 5 redirects and checks for a 2xx or 3xx answer. It can also look for a keyword in the page.
HTTP Advanced
Choose the method, headers, body and authentication, which status codes count as up, and what the response must or must not contain. A slow answer can mark it degraded, and it can warn before the certificate expires.
ICMP ping
Pings a host name or address over IPv4 or IPv6. For servers, routers and anything else that has no web page.
TCP port
Opens a connection to a port and checks that it is accepted. For databases, SSH, game servers and other services that do not speak HTTP.
SSL certificate
Warns before a certificate runs out, 14 days ahead by default, and goes down once it has expired or the TLS handshake fails.
Domain expiry
Looks the domain up with its registry once a day and warns before the registration lapses. A second probe confirms that a domain is expiring, has expired or was renewed before you are told.
IMAP
Connects to your incoming mail server and checks that it answers a command, without logging in. With TLS on, it checks the handshake too and can warn before the certificate expires.
SMTP
Connects to your outgoing mail server and checks that it answers a command, without sending mail. With TLS on, it checks the handshake and the certificate as IMAP does.
Heartbeat
For backups, cron jobs and other scheduled tasks. The job requests its ping URL each time it finishes. When no ping arrives within the period plus the grace, or the job requests the URL with /fail on the end, the monitor goes down.
Checked twice before anyone is told.
A network route can fail between one of our locations and your server while everyone else reaches it fine. So one location is never enough to open an incident.
- A check fails at one location.It is recorded, and nothing is sent.
- Another location checks again.Within seconds, not at the next interval.
- Both failed: the incident opens.Its contacts are alerted, and the failing locations save a network path trace.
- It passes again: the incident resolves.The contacts hear that too, with how long it was down.
Where the checks run.
Three locations today, all in the EU. Each checks over IPv4 and IPv6. More regions are coming, and they go on this list when they are live.
| Location | Country | Region | Code | IPv6 |
|---|---|---|---|---|
| Falkenstein | Germany (DE) | Europe | fsn1 | Yes |
| Nuremberg | Germany (DE) | Europe | nbg1 | Yes |
| Helsinki | Finland (FI) | Europe | hel1 | Yes |
Alert channels.
Add a contact once and attach it to the monitors it should hear about. Acknowledge an incident and its reminders stop.
Email
To any address, verified once.
SMS
Paid from your plan's monthly SMS credits.
Phone call
To a mobile or a landline. The call reads out the monitor, what failed and how many locations see it.
Pushover
To your phones and desktops, with your Pushover user key.
Discord
To a channel, through a Discord webhook.
PagerDuty
Opens and resolves PagerDuty incidents, so your rota applies.
Slack
One message per alert to a channel, with a link to the incident.
Microsoft Teams
A card per alert in a channel, through a Teams workflow link, verified by a code we post there.
Webhook
JSON to your own HTTPS endpoint, once a DNS TXT record proves the domain is yours.
The rest of it.
Maintenance windows
Hold alerts during planned work, for all monitors or the ones you pick, for up to 7 days. When the window ends, only what is still down alerts.
Network path traces
When a monitor goes down, the failing locations run an MTR to the target, so the incident shows where the path stopped.
REST and GraphQL API
Each API token has a role, viewer, member or admin. The REST side has an OpenAPI document.
MCP server
Connect Claude or another AI assistant to read and manage your monitors and incidents.
Terraform, OpenTofu and Pulumi
Providers for all three describe monitors and contacts as code. Changes made in the web app show up in the plan.
Import
Bring monitors and contacts over from UptimeRobot, HetrixTools or NodePing. You review everything before it is created.
Teams
Invite people to your organisation with a role. The audit log records who changed what.
Sign-in
Passkeys, or a password with a second factor from an authenticator app.
Dark theme
For the dashboard and for this site. It follows your system unless you pick one.
Questions about checks.
When does a monitor count as down?
When a check fails at one location and a check from another location fails too. A single failure from one place is recorded, but it opens no incident and sends no alert.
What does NO DATA mean?
That no results came in for a while, for example while a location is being replaced. It is shown as its own state and is never counted as an outage.
Can I choose the locations?
Yes. Pick the locations or the region for each monitor. Every monitor needs at least two, so that a failure can be confirmed.
Do heartbeats need a location?
No. We check nothing from our side, because your job calls us. GET, HEAD and POST requests to the ping URL all count.
Just monitor it.
Add a URL, pick how often to check it, choose who hears about it.
Start your free trial