What it is
A ping monitor sends several echo probes to a host you choose and reports the average round-trip latency and the percentage that got no response. Regions, alerting, and history all work exactly as they do for every other check type: the check runs from all four regions (or the subset you pick) on your chosen interval, and a state change fires the same alert fan-out.
The transport, honestly
This is not raw ICMP. A genuine ICMP echo needs a privilege (a raw socket or an unprivileged "ping socket") that our cloud probes, running as an unprivileged process on Fly Machines, are not granted -- and Node has no built-in way to speak ICMP even when that privilege exists. Rather than pretend otherwise, a ping check measures reachability the way our TCP monitors already do: it connects to whichever of a host's commonly-open ports (443, then 80, then 22) accepts a connection, and times that. The first probe in a run finds the port; every following probe in that same run reuses it, so the latency series is a real round-trip measurement against one endpoint, not a fresh scan each time.
The practical effect: a host that answers real ICMP pings but has all three of those ports closed or firewalled reports as down here. For that narrow case, a TCP check against the specific port you know is open remains the more precise tool.
What counts as up
At least one of the probes in a run got a response. Packet loss is still recorded and shown even on an otherwise-up check, so a flaky path shows up in the numbers before it is bad enough to open an incident. Two consecutive down runs are required before an incident opens, the same confirmation rule every other monitor type uses.
Choosing a host
Enter a bare hostname or IP address, with no scheme, port, or path. Targets have to be publicly reachable from our cloud probes: private, loopback, link local, and cloud metadata addresses are refused, both when you create the monitor and again every time it runs. To watch a host behind your firewall, bind the check to an installed Monitor agent instead -- it runs the same measurement from inside your network.
Availability and pricing
Ping monitors are available on every plan, free included, and count against the same monitor-slot limit as every other check type. There is no separate allowance and no extra cost. See pricing for what each plan includes.