Monitor the services that are not websites
Port monitoring checks that a service is accepting connections. StatusLoom opens a TCP connection to the hostname and port you choose — a database on 5432, a mail server on 587, a game server on 25565 — closes it straight away, and emails you when the connection fails.
What is TCP port monitoring?
Every network service listens on a numbered port: websites on 443, SSH on 22, PostgreSQL on 5432, SMTP mail on 25 and 587. When the service behind a port stops, the website on the same server can look perfectly healthy.
A port check knocks on that one door. If the connection opens, something is listening. If it is refused or times out, the service is down — and StatusLoom tells you which of the two it was, because they point at different fixes.
When you need it
Real failures this catches — each one invisible to an ordinary uptime check.
The mail server that did not come back
An automatic update rebooted the server and the mail service failed to start. The website was fine; quotes and invoices silently stopped going out. A check on port 587 catches it in two checks.
A firewall change that cut off the database
A cleanup of cloud firewall rules removed the database port. The check times out instead of being refused — which points straight at the network, not the database.
A game server for a community
A mod crashes the server at 11 PM. The owner hears about it from a port check, not from a flood of "server down??" messages.
Slow before it is down
Connection time is recorded on every check and charted. A server that goes from 40 ms to 400 ms often shows trouble a day before it fails.
How it works
- Step 01
Add a port monitor
Choose Port, then enter the hostname or IP address and the port number. No http://, no path — just the host.
- Step 02
We connect and close
On your interval — every minute to every hour — we open a TCP connection, time it, and close it. Nothing is sent, nothing is logged into.
- Step 03
Refused and timed out, told apart
Refused means the machine answered but nothing is listening: a stopped service. Timed out means nothing answered: a firewall or a dead server.
- Step 04
Two failures in a row, one email
The alert names the host, the port and which kind of failure it was. The recovery email follows the first successful connection.
A real failure, step by step
A company mail server, mail.example.in:587, on a five-minute interval.
09:00
Port open
Connected in 45 ms.
09:12
Server restarts after an update
The mail service fails to start.
09:15
First failure
Connection refused. Recorded, nothing sent.
09:20
Second failure — outage declared
Email: mail.example.in:587 — connection refused.
09:35
Recovered
Service started, port open again. One recovery email.
Common ports
Any TCP port from 1 to 65535 can be checked. These are the ones people most often watch.
| 22 | SSH |
|---|---|
| 25 / 465 / 587 | Email (SMTP) |
| 3306 | MySQL |
| 5432 | PostgreSQL |
| 6379 | Redis |
| 27017 | MongoDB |
| 25565 | Minecraft |
What it does not do
Stated plainly, so nothing surprises you later.
- Public addresses only. Private and internal ranges — 10.x, 192.168.x, 127.0.0.1, cloud metadata addresses — are refused, and re-checked on every connection in case DNS changes later. To check a database this way, allow our server through your firewall rather than opening the port to everyone.
- An open port means something is listening, not that it is healthy: a database that accepts connections but fails every query still passes.
- TCP only. UDP services cannot be checked this way.
- Checks run every minute at most. Anything faster turns a monitor into a connection flood aimed at your own server.
Common questions
- Does the port check log in to my database or send any data?
- No. It opens a TCP connection and closes it immediately. Nothing is sent, so nothing is authenticated or queried.
- What is the difference between "connection refused" and "timed out"?
- Refused means the server answered but nothing was listening on that port — usually a stopped service. Timed out means nothing answered at all — usually a firewall dropping the traffic, or the server being off.
- Can I monitor a server on my private network?
- No. Only publicly reachable addresses can be monitored, and private ranges are refused on every check. This stops the service from being used to scan internal networks — including ours.
- Is response time recorded for port checks?
- Yes. The time to open each connection is recorded and charted like an HTTP response time, so a slowing server is visible before it fails.
- Is port monitoring free?
- Yes. A port monitor counts as one of the five monitors on the free plan.
Other ways to monitor
- Website monitoring →Status codes, timeouts, DNS and certificate failures, every minute.
- Keyword monitoring →Alerts when a page answers 200 OK but is missing the text it should show — or shows text it should not.
- Heartbeat monitoring →Your cron job or script pings a URL when it finishes. If the ping does not arrive on time, you get an email.
