Why a Continuous Ping Beats a One-Off Test
The most frustrating connection problems are intermittent. The line is perfect for ten minutes, then a call freezes for two seconds, then it is perfect again. A speed test run after the fact shows everything is fine, because the problem was gone by the time you tested. A continuous ping solves this by sending one packet every second and printing the result, so it is already running when the drop happens and records it with a timestamp. You are no longer trying to reproduce a glitch on demand — you set a trap and wait for it to spring.
The Commands
The key difference between platforms: on macOS and Linux ping runs continuously by default until you stop it, while on Windows you must add -t.
# Windows — continuous ping (Ctrl+C to stop, prints running stats)
ping -t 1.1.1.1
# macOS / Linux — continuous by default (Ctrl+C to stop)
ping 1.1.1.1
# Slow it down or speed it up (interval in seconds)
ping -i 0.5 1.1.1.1 # macOS/Linux: every half second
# Ping your router gateway first to test the local link
ping -t 192.168.1.1 # Windows
ping 192.168.1.1 # macOS/LinuxWhen you press Ctrl+C, the test prints a summary: packets sent, received, percent lost, and the minimum/average/maximum round-trip time. That summary line is the headline result; the per-line output is where you find the exact moment things went wrong.
Logging It to a File With Timestamps
To leave a ping running for hours and review it later — or to hand evidence to your ISP — write it to a file. A raw log without timestamps is hard to correlate with "the call dropped around 8:15," so add them.
# Windows — log to a file
ping -t 1.1.1.1 > ping-log.txt
# macOS / Linux — prepend a timestamp to every line
ping 1.1.1.1 | while read line; do echo "$(date '+%H:%M:%S') $line"; done | tee ping-log.txt
# Linux alternative with built-in timestamps
ping -D 1.1.1.1 | tee ping-log.txtLeave it running through the window when the problem usually appears, then open the log and search for the word timeout (Windows) or lines showing an unusually high time= value. Each one is timestamped, so you can line them up against when you noticed the stutter.
Reading the Output
| What a line shows | Meaning |
|---|---|
time=12 ms steady, line after line | Healthy, stable connection |
time=12 ms then a sudden time=240 ms | A latency spike — likely congestion or bufferbloat at that instant |
Request timed out / no reply | A dropped packet — the reply never came back in time |
| Several timeouts in a row, then recovery | A brief outage or a flapping link / Wi-Fi disconnect |
| Summary shows "2% packet loss" | 2 of every 100 packets lost — enough to disrupt calls and games |
For a deeper read of what each number in the summary means — and what counts as a good average and jitter — see how to read ping output.
The Two-Target Method: Local vs Upstream
Run two continuous pings side by side in separate windows — one to your router gateway (192.168.1.1) and one to a public IP (1.1.1.1). Then compare them when a drop occurs:
- Gateway drops too: the problem is inside your home — Wi-Fi interference, a weak signal, or a failing cable. Move to a wired connection to confirm.
- Gateway stays clean, public IP drops: your home network is fine; the loss is upstream at the modem, the line, or the ISP. This is the evidence you bring to a support call.
If the loss is upstream, the natural next step is a per-hop view that shows where on the path it starts — that is what MTR and a packet-loss test are for.
A Note on ICMP De-prioritisation
Ping uses ICMP, and some routers along the path treat ICMP as low priority — replying slowly or not at all when busy, even while they forward your real traffic perfectly. This is why a single intermediate hop can show loss in a continuous ping while your actual browsing is fine. Trust loss that appears at your final destination and persists; be skeptical of loss that appears at one middle hop but clears at the hops beyond it.
Frequently Asked Questions
How long should I run a continuous ping test?
Long enough to capture the problem at least a few times. For a drop that happens every few minutes, 15–30 minutes is enough. For an intermittent issue that appears once or twice an hour, run it for several hours — ideally spanning the period when the problem usually occurs. The whole point of a continuous test is to be running when the rare event happens.
What address should I ping?
Ping in stages. First ping your router gateway (192.168.1.1) to test the local link; clean here means your LAN and Wi-Fi are fine. Then ping a reliable public IP such as 1.1.1.1 or 8.8.8.8 to test the path to the internet. Comparing the two tells you whether drops originate inside your home or upstream at the ISP.
Why does ping work but show occasional "Request timed out"?
A timeout means a single packet got no reply within the wait window — that is a dropped packet. Occasional isolated timeouts under heavy load can be normal, but repeated timeouts, or timeouts that line up with stutter in calls and games, indicate real packet loss worth diagnosing further with a per-hop tool like MTR.
Does a continuous ping use a lot of data?
No. Each ping is a tiny packet — roughly 64 bytes out and back. Even running for hours at one packet per second it consumes a negligible amount of data, far less than loading a single web page. It will not affect your usage or interfere with other traffic.