What is a port?
A port is a numbered entry point that separates the different services running on one IP address. If the address is the building, the port is the apartment number.
Web traffic uses 80 and 443, SSH uses 22, RADIUS uses 1812. A service being up means it accepts connections on its own port.
- 1
Ping: is the address answering?
- 2
TCP port: is the service listening?
- 3
REST / application: is the response correct?
When the three results differ, they point to the layer where the problem sits.
What a TCP port check tells you
A TCP port check attempts a connection to a specific port on the target. If the connection succeeds, the service is listening; if it fails, the service has stopped, a firewall is blocking it, or the device is unreachable.
This answers the question ping cannot. Ping shows the operating system is up; the port check shows the service on top of it is actually serving.
Can the port be open while the service is still broken?
It can, and it happens often. A service may accept connections while returning an error to every request because it cannot reach its database. A port check will not catch that.
So checks are layered on critical services: the port check is a fast, cheap first signal, while an application-level REST check verifies the response the user actually receives.
- Ping fine, port closed: the service may have stopped
- Port open, REST failing: an application or dependency problem
- All failing: an outage at device or link level
- Intermittent failures: capacity, timeouts or packet loss
Port scanning is not port checking
Port scanning discovers which ports are open on a device and helps build an inventory. A port check is the repeated verification of one known port.
Sentry lists candidates during discovery, and the monitoring template decides which port is checked how often. Security auditing and vulnerability detection are out of scope.