Diagnosing Internet Uplink Issues with Path Ping and MTU Testing

Public DNS resolvers (1.1.1.1, 8.8.8.8) polled every cycle just like any device on the fleet, linked off your NOC routers, turn an internet-wide outage into an immediate, obvious alert instead of a wave of individual customer complaints that take longer to add up to the same conclusion.

Path ping runs from a specific device — not from your own machine — to any internet target, a gateway router, or a custom IP you type in. That distinction matters: pinging from your laptop tells you about your own connection; pinging from the actual site router tells you about theirs.

Payload size is selectable — 56 B, 512 B, 1500 B (standard MTU), or 9000 B (jumbo) — and this is the detail most troubleshooting skips. A link that pings clean at 56 bytes but drops or fragments badly at 1500 bytes is pointing at an MTU mismatch somewhere in the path, not a signal problem — classic symptom: small pages load fine, large downloads or video calls stall.

• Clean at all sizes, high latency — likely a real upstream congestion or routing issue, not local to the site.
• Clean at small sizes, bad at 1500B+ — check MTU configuration along the path (PPPoE links are a classic culprit).
• Bad even at 56B from one specific site — this is local: check that site's backhaul or wireless link, not the internet uplink.

Once path ping narrows down where the problem sits, a device-to-device bandwidth test (10–1000 MB, upload or download, TCP or UDP) gives you an actual throughput number with a live animated Mbps read — useful for confirming a suspected bottleneck is real before you escalate it or roll a truck.

If path ping and the bandwidth test both look clean but customers on one sector are still unhappy, the problem is probably on the wireless side, not the uplink — check the Spectrum tab for interference or the per-AP client roster for a specific struggling client before assuming it's the internet connection at all.