Monitoring

Monitor Network Ports

Check whether a service is reachable and alert when the result changes.

Choose what you are measuring

A TCP or UDP port is a service endpoint, such as HTTPS on TCP 443. A physical firewall or switch port is an interface. Use a TCP or UDP check for the former; use SNMP, a device API or a mapped trap for interface link state.

Pick the observation point

For a private address, enroll a Site Collector on the network and select it under Connect Through. To test public exposure, use a separate authorized collector outside that network. A successful internal check says nothing about whether an Internet client can connect. Hosted probes depend on your service access; a customer does not install or administer the Sysavia panel.

Configure a TCP check

In Design, select the equipment and choose Add Check. Name it, select TCP Port, enter the Host / IP and Port, then select the collector. Start with a 15-second interval. Save and confirm the result before adding an automation.

Up means the TCP connection was accepted. Down means the connection failed, was refused or timed out. Read the result detail. A timeout can mean filtering, a broken route or an unreachable host; it does not prove the port is closed.

Alert when access disappears

Create an automation with When set to Check / Service Is Down. Select the system and TCP check, then add Send Alert. Use two or three matching cycles for brief interruptions and a cooldown to avoid repeated noise. For a 15-second check, three samples take roughly 30–45 seconds to establish a sustained failure.

Alert when a port becomes reachable

Create a separate rule using Check / Service Is Up against the same check. Name the message for your intent: “HTTPS reachable again” for recovery, or “SSH is reachable from the outside collector” for unexpected exposure. These are state-based rules: the first matching observation can trigger an alert, even if no earlier transition was recorded. Check cooldown and rearm behavior when testing.

Send Alert writes to the organization’s Activity feed. Use a configured HTTP or API action when you need delivery to an external notification service.

Test the application as well

TCP 443 can accept connections while the web application is failing. Add HTTP Health for an HTTP response check, or an API check that validates a specific response value. Keep network reachability and application health as separate observations.

UDP needs a real response

Choose UDP Response and supply a request payload appropriate to the service. The current check accepts a text payload and optional expected response text. No reply is a failed check, not proof of a closed UDP port. Services that need binary protocol messages require a suitable agent script or protocol-specific integration.

Physical interfaces and firewall rules

For an Ethernet interface, collect operational link status through SNMP or the vendor API and associate that check with the modeled port. For a firewall rule being enabled or disabled, monitor the firewall’s configuration, API or event log. A successful port probe alone cannot tell you which rule changed.

Watch a known port or discover a new one

The built-in checks watch the endpoints you configure. They are not a continuous scan of every listening port. To detect a new listener anywhere on a server, run an approved agent script that compares its listening sockets with an allowed baseline and reports the result as a check or custom event. An external check is still needed to establish whether that listener is reachable through the firewall.