Ping

Overview

The Ping monitor verifies a server or network device is reachable. It’s the simplest and lightest monitor type.

Settings

Setting	Description	Default
Host	IP address or hostname	—
Interval	Check frequency	shortest your plan allows
Timeout	Fixed check budget (not configurable)	5s
Regions	Where the check runs from	one, per your plan

A URL pasted into the host field works too — the scheme, port and path are stripped before the check runs.

How it works

Each check sends one ICMP echo and measures the round trip.

Because a great many hosts drop ICMP at the firewall while happily serving traffic, a silent host is not immediately declared down: the agent then tries a plain TCP connect on port 443, then 80. If either opens, the host counts as reachable and that connect time is the recorded latency. The check fails only when the ICMP echo goes unanswered and neither port accepts a connection within the 5-second budget.

The agent uses an unprivileged ICMP datagram socket (no CAP_NET_RAW required). Where the host doesn’t allow one at all, the TCP probe is used on its own — the monitor keeps working, it just measures a connect instead of an echo.

Use cases

The Ping monitor is ideal for:

• Dedicated servers — Verify your physical or virtual machines are reachable.
• Network equipment — Monitor routers, switches, and firewalls.
• Basic connectivity — Test network reachability before checking application-level services.

For a specific port other than 443/80, use a TCP monitor instead — it tests the port you name rather than treating it as a fallback.

Metrics collected

Every check records the latency (ICMP round trip, or the TCP connect time when the fallback was used), the region it ran from, and the error when it failed. There is no packet-count, packet-loss or jitter measurement: one echo per check either comes back or does not.

Private hosts are refused by the public regions — an address that resolves inside a private range is only monitorable from a private agent.