The looking glass runs network tests from LayerOne toward an address you choose. No account needed.
Tools
- Ping, four packets, reporting min, average and max latency plus loss
- Traceroute, the hop-by-hop path, capped at a couple of dozen hops
- MTR report, traceroute and ping combined, which is what you want for intermittent trouble
Rate limited to 3 runs per minute per source.
What it is good for
Answering "is the problem the network between us, or my server?" Before opening a ticket about latency or packet loss, run an MTR both ways: from the looking glass to your address, and from your server to your own location. Attach both.
# From your server, toward wherever your users are
mtr --report --report-cycles 50 198.51.100.1
Loss that appears at one hop and then disappears is a router deprioritising ICMP, not a fault. Loss that appears at a hop and continues to the destination is real.
What it is not
It tests the path, not your server. If your service is down but the address pings, the problem is inside the guest: use the browser console.