PINGFLEET Sign in

Check locations

Multi-location checks with consensus

Where a check runs from decides what it can tell you. One machine failing to reach you is not an outage.

Per monitor

Which location saw what

Every check records the location that made it, so a monitor can say it is reachable from Frankfurt and not from New York rather than averaging the two into a shrug.

An HTTP monitor in PingFleet showing uptime, average latency and a bar per check.

The fleet

The probe fleet

Checks run from separate providers on separate autonomous systems, so that two of our machines cannot share one network's bad day. Every address is published at /ips.json, as a file you can paste into a firewall rule.

Frankfurt de-1 · AS63949 Strasbourg fr-1 · AS16276 New York us-1 · AS14061 Bacau ro-2 · AS39383

Marked points are the machines that run checks. The arcs are checks leaving them for targets elsewhere, not places we check from.

Consensus

Consensus between locations

de-1 · AS63949 fr-1 · AS16276 us-1 · AS14061 agreement majority of reporting probes up, nobody is woken degraded, shown but not paged down, the escalation starts

Independent networks, not just distant cities

Two servers three thousand kilometres apart at the same provider share that provider's routing, peering and bad afternoons. Geographically they look independent; operationally they are one machine with extra latency.

PingFleet picks locations for network independence first: different hosting companies, different autonomous systems. Distance is the tie-breaker, not the criterion.

Every address is published at /ips.json and /ips.txt, as a file, because the person adding them to a firewall wants a list they can paste into a script.

Agreement before anyone is paged

A monitor is called down when a majority of the locations that reported agree. A single bad route no longer pages anybody.

A location that has never reported is not counted as a vote against, and a vote older than a few intervals stops counting.

The third state

The degraded state

Reachable from Frankfurt, not from New York, named on the monitor. It does not page anybody at three in the morning; it is not hidden either.

Before reporting a partial failure, PingFleet checks whether that location is failing the same way for everybody. If it is, the problem is ours, and you are not told your service is degraded because one of our machines is having a bad hour.

Questions

Can I choose the locations?
On paid plans, yes: either let PingFleet spread them for maximum independence, or pick them explicitly.
How many locations do I get?
1 on the free plan, 2 on Plus and 3 on Pro. Where there is more than one, they have to agree before a monitor is called down.
What happens if a location is retired?
The account is moved to the nearest equivalent and told, because a move shifts the latency baseline and an unexplained step in a graph reads as your infrastructure slowing down.
20 monitors free. No card. Start free