Uptime
A real request to your URL, from outside your network.
Alerts when: Status code outside the range you set, a timeout, a refused connection, a DNS failure or a TLS error.
Each one answers a different question about a site. Most people start with the first and add the rest as they discover what actually breaks.
A real request to your URL, from outside your network.
Alerts when: Status code outside the range you set, a timeout, a refused connection, a DNS failure or a TLS error.
Loads the page and looks for text that must - or must not - be there.
Alerts when: Catches the case a status check cannot: a page returning 200 OK while showing a database error, a blank screen or a suspension notice.
Opens a TLS connection and reads the certificate itself.
Alerts when: Expiry inside three days, a name that does not cover the host, a self-signed certificate or a broken chain. Warnings start 14 days out.
Asks the registry directly over RDAP, falling back to WHOIS.
Alerts when: Registration ending within three days. Warnings start 30 days out, which is enough time to renew calmly.
Resolves A, AAAA, MX, NS, TXT or CNAME and compares them with what you expect.
Alerts when: A record disappearing or changing - which catches both a botched migration and a hijacked domain.
Connects to a specific port: mail, database, SSH, anything listening.
Alerts when: Nothing accepting connections, or a connection that times out.
Three ICMP echo requests to the host. Where ICMP cannot be sent, a TCP handshake on 443 or 80 stands in, and every result records which of the two answered.
Alerts when: Nothing comes back by either route. Useful for servers that answer no HTTP at all - though a host behind a firewall that drops ICMP will be measured over TCP instead.
Works the other way round: your job calls us when it finishes.
Alerts when: The call not arriving in time. The only way to watch a nightly backup or a cron task with no public address.
Submits the form the way a visitor would, with the fields you give us, and reads the reply.
Alerts when: The success message not coming back. A form that quietly swallows enquiries is the most expensive kind of broken, because the site looks perfect.
Checks whether a page is telling search engines to stay away.
Alerts when: A stray noindex header or meta tag, or a robots rule - usually left behind after a staging site goes live. Nothing else on this list would notice.
Reads the DNS records your mail depends on: MX, SPF, DMARC and the DKIM selectors you name.
Alerts when: A record that has gone missing or been edited, which is how a domain starts landing in spam. This checks the records, not the mail server itself.
One failed request never raises an alert. We re-check 30 seconds later, and only a confirmed failure counts. Before believing any failure we also verify our own connectivity against reference hosts - and if more than a third of everything we watch fails at once, we assume the fault is ours, hold every alert and wake our own operators instead.
Email, Telegram and signed webhooks. Quiet hours hold back certificate and domain reminders, but never an outage. A monitor flapping up and down is collapsed into one message instead of fifty.
A public status page, on your client's own domain if you want, showing the services you choose under names you choose. They stop emailing you to ask whether the site is down.
Uptime history, incident records with causes and durations, and reports you can schedule straight to the client on the first of every month.
No card, and commercial use is allowed.
Start monitoring