How ping checks work

A ping check is the simplest probe in monitoring. The monitor sends an ICMP echo request — a tiny packet whose only job is to ask "are you there?" — and the target's operating system answers with an echo reply. The check measures two things: whether a reply came back at all, and how long the round trip took in milliseconds.

What a successful ping proves

The host is powered on and its network stack is running.
A network path exists to the host and back — routing, cabling, and every hop in between are doing their job.
The round-trip time gives you a clean read of network latency toward that host.

Because the reply comes from the operating system kernel, ping works against anything with an IP address — no service needs to be installed or configured on the target.

What it doesn't prove

Host up is not service up. The kernel answers the echo whether or not the applications above it are alive. A server whose web service has crashed, whose database is refusing connections, or whose disk is full will keep answering ping cheerfully. The rule of thumb: ping tells you the machine is reachable; only an HTTP check or a TCP check tells you a service is answering. If your client depends on a specific service, monitor that service directly and use ping for the layer underneath.

When ping is the right check

Network devices — switches, routers, firewalls, access points. They run no website or database you could probe, but they answer ICMP, so ping is often the only practical check.
Internet links — ping the router or CPE address of a branch to watch the link's stability and latency.
Base-layer liveness — a ping monitor next to your service monitors helps you tell "the whole machine is gone" apart from "one service died on a healthy machine".

Firewalls that drop ICMP

Many firewalls block or rate-limit ICMP by policy. A host filtered this way looks permanently down to a ping check while working perfectly for its users. Some networks also deprioritize ICMP under load, which makes ping latency jumpier than what real traffic experiences. So before saving a ping monitor, press Test connection: if the target is known-healthy but the probe fails, ICMP is probably being filtered — switch to a TCP check against a port you know is open.

Ping in UpAll

Choose the Ping check type and give a hostname or IP address as the target. The timeout is fixed at 10 seconds; a reply that never arrives, or arrives too late, counts as a failed check, and an incident opens after the configured number of consecutive failures (default 3 — see false positives and flapping). Latency above the degraded threshold (default 100 ms for ping, capped at 1999 ms) marks the check degraded without opening an incident. Setup details: Ping monitors.