30 s buckets · in GAME TRAFFIC mode this is true UDP loss; in STANDARD mode it's the failed-probe ratio (TCP retransmits hide some loss)
spot patterns — e.g. congestion every evening
Saturates your download for ~8 seconds while latency probes keep running, then compares your latency idle vs under load. A big increase means your router queues packets — a hidden lagspike source whenever someone on your network starts a download or stream. Uses roughly 10–30 MB of data.
| Time | Severity | Peak latency | Duration | Above baseline |
|---|---|---|---|---|
| no lagspikes or outages recorded yet — looking good | ||||
| Start | Duration | Status |
|---|---|---|
| no outages recorded — connection has been up | ||
Everything this monitor has recorded — lagspikes, outages, jitter, packet loss, uptime — bundled into proof you can hand to your internet provider. The report is print-friendly and written in plain language; the CSV has every event with ISO timestamps for your own records.
Bandwidth and stability are two different things. A speed test measures how much data your line can move in a few seconds — the size of the pipe. Online games don't need a big pipe; they need a steady one. A game sends small updates dozens of times per second, and what ruins your match is not slow throughput but moments where those updates arrive late or not at all:
Ordinary speed tests average all of this away. LagScope probes your connection every 500 ms and logs every spike, dropout and outage with an exact timestamp — so you can see the problems your speed test hides, and prove when they happen.
"Looks fine on our end" is what every gamer hears. ISPs check the signal at your modem at one moment in time — they never see the 30-second outage you had at 21:40 last night. The way to win that conversation is with timestamped evidence:
A report showing "23 lagspikes (4 severe) and 2 outages totalling 3 min 12 s of downtime between 14:02 and 18:47" is much harder to dismiss than "my game lags". Run it during the hours you usually play — patterns like congestion every evening are exactly what ISPs can act on.
A lagspike is a short moment where your latency jumps far above its normal level — for example from 25 ms to 300 ms for a second or two. Even brief spikes cause missed shots, freezes and rubber-banding in online games, and they are invisible to ordinary speed tests.
Speed tests measure maximum bandwidth over a few seconds and report an average. Lagspikes, jitter and short dropouts happen between those averages — a line can deliver 500 Mbps and still stutter every few seconds. LagScope probes your connection continuously (every 500 ms by default) so these short events are recorded with exact timestamps.
Jitter is the variation between consecutive latency measurements. Under 5 ms is excellent, 5–15 ms is usually fine for casual play, and anything consistently above 15–20 ms causes visible stutter, delayed actions and robotic voice in calls — even when average ping looks good.
Bufferbloat is excessive queuing in your router or modem: when the line gets busy (someone starts a download or stream), packets pile up in a buffer and latency balloons. It is one of the most common causes of lag that only appears while the connection is under load. LagScope's stress test measures exactly this by comparing your latency idle vs under load.
No. Latency probes are effectively zero-bandwidth (an empty 204 response every 500 ms), and the periodic throughput test uses about 96 KB per run — roughly 45 kbps on the default setting, and you can turn it off. Only the on-demand bufferbloat stress test uses more (about 10–30 MB), and it runs only when you click the button.
Everything stays in your browser. Lagspike events, outages and settings are kept in localStorage on your device — nothing is uploaded or stored on a server. The ISP evidence report and CSV export are generated locally from that data. Two optional exceptions, both off by default: if you paste your own Discord webhook URL in Settings, alert messages about severe lagspikes and outages are posted to that Discord channel; and if you enable "Contribute anonymous stats", only your ISP name, country and connection metrics (never your IP) are sent to the public ISP leaderboard.
Yes — paste a Discord webhook URL in Settings and LagScope posts a message to that channel whenever a severe lagspike hits or an outage starts and ends (with the downtime). Alerts are fire-and-forget: a failing webhook never interferes with measurement.
Standard mode measures with small HTTP requests over TCP — and TCP silently retransmits lost packets, which hides true packet loss and stretches spike durations. Real games send small UDP packets 20–60 times a second with no retransmission. GAME TRAFFIC mode reproduces exactly that: it opens a WebRTC DataChannel in unreliable/unordered mode and sends 128-byte packets at 30 Hz — genuine UDP on the wire. That gives you true packet loss (a packet that never comes back within 1 s is lost, like in a game) and latency numbers that match what your game actually experiences. If the UDP path can't be established, LagScope automatically falls back to standard HTTP probing and tells you.
Every measurement is a path between two points: your connection and the LagScope server. The header shows both ends — your ISP and city (looked up from your public IP, never stored) on the left, and the server's location on the right. The MEASURING VIA line names the node your probes actually go to (see the next question). A great result means the path between these two points is healthy; it is a good real-world proxy for the route to nearby game servers, but routes to other destinations can differ.
LagScope measures to your fastest node (US or EU, chosen automatically by real measured latency at page load). Both nodes are probed when the page opens and the one with the lower median round-trip time is used for the session — including GAME TRAFFIC (WebRTC/UDP) mode. The choice is re-made on every page load because routing changes; you can also pin a node in Settings under "Measurement node".
Check the LagScope leaderboard — it ranks ISPs by real, anonymized measurements (gaming-readiness score, latency, jitter, packet loss and lagspikes per hour) contributed by LagScope users who opt in, grouped by ISP and country. No IP addresses or personal data are collected.