How do I know the ping of a proxy server?

To know the ping of a proxy server, time a real request through the proxy instead of sending ICMP packets to its IP address. The quickest way is one curl command with the proxy login and password and the write-out timers, which prints the TCP connect time to the proxy and the time to the first byte from the target site. Run it twenty times, write down the median and the worst result, and you have the two numbers that describe the ping for your task.

The question about proxy ping is asked by clients who need connection speed. Most often they are traders on stock exchanges, parsing specialists or users who need a stable connection to automate processes.

Ping requirements depend on the tasks: for some clients a delay of 100 ms is acceptable, for others even 20 ms is critical. The usual ping command helps neither group, because a rented private proxy answers only on its port and only after authorization, while ICMP replies come from the host itself. Below are the factors that make up those milliseconds, the values our proxies show in practice and four ways to measure the ping when the proxy requires a login and password.

What is ping and why is it important?

Ping is measured in milliseconds and shows how long it takes for data to travel from the device to the server and back. For example, if you send a request through a proxy to access a website, the ping shows how long it takes to get a response. It matters because every new connection needs several such round trips before the first byte of a page arrives, so a proxy with a 100 ms ping adds well over 300 ms to each fresh request.

For some people, ping is critical. For example, if tasks require instant response, such as scalping on the stock exchange, working with captchas or online games. And if you are engaged in parsing or multi-registration of accounts, a delay of 50 to 100 ms often does not affect the overall result.

Through a proxy the route has two legs, from you to the proxy and from the proxy to the site. The answer returns the same way, so the ping you feel is the sum of both. People notice a delay of 100 to 200 milliseconds and call anything above 300 milliseconds sluggish, as the perception data in High Performance Browser Networking by Ilya Grigorik shows. Grigorik, who wrote the book while working on web performance at Google, is blunt: "Latency, not bandwidth, is the performance bottleneck for most websites!" A proxy cannot give you bandwidth, but it can easily double your latency.

What does the ping depend on?

The ping of a proxy is built from distance, network type, server load and your own connection, roughly in that order of weight. Distance sets the floor. A signal in optical fiber travels at about two thirds of the speed of light, no server can make that leg shorter, and the other three factors only add milliseconds on top of it.

Rarely do the four act alone. A datacenter proxy 300 km away on an idle server answers in about 10 ms. The same proxy at peak hour on a shared line may take 40 ms, and a mobile proxy in your own city starts at 50 ms before any load appears. When a measurement jumps, the useful question is which leg changed. The answer decides what to do next, whether to move to a closer server, to an individual line or simply to a cable instead of Wi-Fi.

Each factor has typical values, and knowing them turns a raw number into a diagnosis. A 15 ms result to a server in your country is normal, 60 ms to the same server means queuing or a bad first hop, and 250 ms from a mobile proxy is ordinary rather than broken. The four subsections below give those reference values together with the measurements they come from.

How does the location of the proxy server affect ping?

The farther the proxy server is from your device or target site, the higher the ping. For example, requests from Europe to a server in the USA will show a delay of 100 to 150 ms, while within the same country from 10 to 30 ms.

Physics sets the floor. Light in optical fiber covers roughly 200,000 kilometers per second, so the 5,585 km straight line from New York to London costs about 56 ms round trip in theory. Real cables are longer than the straight line, as the latency primer in High Performance Browser Networking shows with a table of routes. Carriers publish what the backbone actually delivers. In July 2026 Verizon measured 70 ms across the Atlantic, 14 ms inside Europe, 34 ms across North America and 112 ms across the Pacific between its backbone routers. Your last mile and the proxy's last mile come on top, which is why a Europe to US request lands at 100 to 150 ms rather than at 70.

Count both legs when choosing a location. The site feels the sum of your path to the proxy and the proxy's path to the site. The first of those legs is repeated on every new connection, so keep it short whenever the task opens many connections.

Why do mobile proxies have a higher ping than server proxies?

Mobile proxies use the networks of cellular operators. They usually have higher ping than server proxies due to the peculiarities of the infrastructure. To compare, a mobile proxy will show ping in the range of 50 to 300 ms, while a server proxy in the same location shows 5 to 50 ms.

Radio is where the extra milliseconds come from. A phone or a 4G modem shares a cell with other subscribers, waits for a transmission slot and hands over between towers, and each of those steps adds delay and jitter. Opensignal's measurements show that even in the leading 5G markets only a part of users consistently get latency below 50 ms with jitter under 10 ms. The same study found that 5G Standalone networks cut latency by 20 to 30 percent compared with the common Non-Standalone setup. A mobile proxy inherits whatever its SIM card gets at that moment.

Residential proxies sit in between. They run on home cable, DSL or fiber lines. The FCC's Measuring Broadband America report puts the idle latency of those lines at 7 to 14 ms for fiber, 12 to 24 ms for cable and 23 to 34 ms for DSL before any distance is added. A server proxy in a datacenter skips the last mile entirely, which is why it starts from 5 ms.

How does server load affect the ping of a proxy?

Servers that process thousands of requests simultaneously will inevitably increase ping due to queues. It is important to choose a provider that limits the number of clients per server and supports load balancing. Load does not change the distance, it adds waiting time, so the symptom is a ping that is fine at night and two or three times higher at peak hours.

Queuing is one of the four components of latency, next to propagation, transmission and processing, and it is the only one that depends on how busy the path is. The FCC measures home connections both idle and under load for exactly this reason, and its report notes that latency under download load is significantly higher than idle latency, most of all on DSL. A proxy server behaves the same way. An idle machine answers in a few milliseconds, while a machine that relays thousands of parallel streams makes every packet wait its turn in the queue.

Shared lines carry this risk more often than individual ones. In our catalog a shared IPv4 address is used by up to three people and an individual address by one, so a neighbor's bulk download on a shared line shows up in your ping. The practical check is to measure at nine in the morning and at nine in the evening and compare the slow tail rather than the average.

Does my own internet connection affect the proxy ping?

The proxy ping you see always includes your local network. If you have an unstable connection, even a proxy with minimal latency will show high numbers. So yes, your Wi-Fi, your router and your provider take part in every measurement, and a bad first hop can add more than the proxy itself.

Two things on your side inflate the result most often. Wi-Fi on a crowded 2.4 GHz channel adds random spikes of 20 to 100 ms, and an upload running in the background fills the router's buffer so that every other packet queues behind it. The FCC report measures home lines idle and under download and upload load precisely because the loaded values differ from the idle ones. A cable between the computer and the router removes the first problem. Pausing cloud backups during a test removes the second.

To separate your share from the proxy's share, measure a nearby reference at the same time. Ping your router and your provider's DNS server, or run a speed test to the closest server, and note the latency. If the reference sits at 30 ms instead of the usual 5 ms, the spike is yours and not the proxy's. mtr on Linux and macOS or tracert on Windows shows the delay hop by hop and points to the first hop where it grows.

What's the ping on our proxies?

On our server proxies the ping within a single country is 5 to 20 ms, and between countries it is 30 to 150 ms depending on the route. Europe to the USA stays near the upper end, because the transatlantic backbone alone takes about 70 ms between carrier routers. On mobile proxies the typical range is 50 to 300 ms, depending on the operator's cell and the time of day.

Inside each range the spread comes from the server you pick and from the time of day. For Russia we offer four server groups, S1 to S4. The way to choose between them is to run the curl loop from the measurement section against each group and keep the lowest median from your location. A proxy that is 5 ms better in the morning and 40 ms worse in the evening loses to a steady 15 ms line.

These ranges describe a wired computer next to its router, so add your own last mile from the previous section to predict what you will see.

How to measure ping on a proxy?

For a proxy that requires a login and password, the ping is measured by timing a TCP connection to its port and a full request through it. The tools for that are curl, a short Python script, a desktop program that supports proxy authorization or a browser with the proxy configured. The usual ping command works only for open hosts, because it sends ICMP packets that the proxy service never sees.

The easiest way to find out the ping is to enter the ping command at the command line and type the IP of the proxy.

But it doesn't work with proxies that are accessible only with login and password. That is, with any rented private proxies.

Even where ICMP is answered, its numbers deserve caution. Orb's analysis of the ping command explains that many routers deprioritize or rate-limit ICMP, so during congestion ping packets are delayed or dropped first while TCP traffic flows. A proxy behind a gateway may also answer ping from a different machine than the one relaying your requests. A TCP connect to the proxy port shows how fast the proxy accepts you, and the time to the first byte through it shows what your software will actually wait.

In this case, there are several other options to check ping. Let's tell you about all of them.

How do I check proxy ping with curl?

You can check the availability and speed of the proxy via the curl tool in the same cmd. Since 2018 curl ships with Windows 10 and 11, so nothing needs installing, and the same command works on macOS and Linux. With the write-out timers one run prints the ping as the TCP connect time to the proxy, the moment the tunnel and the TLS handshake are done, and the time to the first byte of the page.

Example command for authorization via proxy:

curl -x [proxy-server]:[port] -U [login]:[password] https://example.com -o /dev/null -s -w "connect %{time_connect}  tunnel+tls %{time_appconnect}  first byte %{time_starttransfer}  total %{time_total}\n"

• -x specifies the proxy.
• -U specifies login and password.
• https://example.com is the address of the target site through which the proxy is checked.
• -w prints the timers. In the curl manual, time_connect is the TCP connection to the proxy, time_appconnect includes the CONNECT tunnel and the TLS handshake with the site, and time_starttransfer is the first byte of the response.
• -o /dev/null -s disables output of page content to the terminal.

The result will show how long it took to connect through the proxy. Run it twenty times in a loop and look at the median and the largest value, because a single run may catch a retransmission. On Linux and macOS wrap it in for i in $(seq 1 20); do ...; done, in PowerShell in 1..20 | % { ... }. The gap between time_starttransfer and time_appconnect is roughly one round trip to the site plus its processing time; when it is large, the proxy is not the bottleneck.

How do I check proxy ping in Python?

In Python the ping of a proxy is measured with the requests module. Send a GET through the proxy with stream=True, time it with time.perf_counter() and repeat twenty times to get a median and a worst case. The stream flag stops the transfer after the headers, so the result is the time to the first byte rather than the time to load the whole page.

If you know Python, you can write a script to check the proxy:

import requests, statistics, time

 

proxy = {

    "http": "http://login:password@proxy-server:port",

    "https": "http://login:password@proxy-server:port",

}

 

samples = []

for _ in range(20):

        start = time.perf_counter()

        try:

            r = requests.get("https://example.com", proxies=proxy, timeout=10, stream=True)

        samples.append((time.perf_counter() - start) * 1000)

            r.close()

        except requests.exceptions.RequestException as e:

        print(f"Error: {e}")

 

if samples:

    print(f"Proxy ping: median {statistics.median(samples):.0f} ms, worst {max(samples):.0f} ms")

This code measures the latency (ping) to the target site through the proxy. Each call opens a new connection, so every sample includes the TCP handshake with the proxy, the CONNECT tunnel and the TLS handshake with the site, which is the worst case your software meets. To see the latency of an already open connection, the one a browser or a scraper reuses, create a requests.Session() and call get() on it. The second and later samples then drop to roughly the round trip of both legs. Keep timeout=10, otherwise a dead proxy hangs the script for minutes. The statistics module is part of the standard library, so requests stays the only external dependency.

Which programs show the ping of a proxy with a login and password?

Desktop programs that support proxy authorization show the request time through a proxy the way your applications will see it. HTTP Debugger Pro, Proxifier and PingPlotter each do it for a different kind of work, and PsPing adds a plain TCP test of the proxy port for Windows users.

HTTP Debugger Pro is a program for diagnosing and testing HTTP/HTTPS requests. Ideal for tasks with proxy servers. You can download a one-week trial version from the official website. It tests a proxy with a login and password, shows request headers, responses and authorization parameters, and records the execution time of every request. That makes it convenient for watching how stable a proxy stays over an hour of real work.

Proxifier configures proxy access for any application and allows connection tests.

PingPlotter can be configured to work through a proxy with login and password to analyze the route. PsPing from Microsoft Sysinternals pings a TCP port directly, for example psping -n 50 host:port, and reports the connect time with 0.01 ms resolution. That covers only the first leg, your path to the proxy, but it needs no credentials and shows at once whether the port is reachable and how stable the path to it is.

Can I check the ping of a proxy through online services?

Yes, but only after the proxy is configured in the browser, and only with an understanding of what the site measures. Speedtest.net run through a proxy picks a test server near the proxy's exit, so the ping it shows is your route to the proxy plus a short hop to that server, not the route to your target site. Ping.eu sends ICMP from its own server to the address you type. Neither number is your route.

Set up a proxy in your browser (e.g. Chrome or Firefox). Go to a website such as speedtest.net or ping.eu.

Still, a speed test makes a decent location check. The site picks its server by the proxy's IP, so a proxy sold as a New York address that gets matched with a server in Frankfurt is registered in Germany. If the ping from Berlin stays near 5 ms, the exit is physically there too, because a detour through New York would add two Atlantic crossings, well over 140 ms.

No website is needed for the most honest browser measurement. Open DevTools in Chrome, load the target page through the proxy and click any request on the Network tab. The Timing tab shows a separate Proxy negotiation phase, and its documentation defines Waiting (TTFB) as one round trip plus the server's processing time. That breakdown tells you whether the delay sits in the proxy, on the route or inside the site.

Why is the first request through a proxy slower than the next ones?

The first request through a proxy is slower because the connection is built in several round trips before any data flows, while later requests on the same connection need one. Over an HTTP proxy the client completes a TCP handshake with the proxy, then sends a CONNECT request and waits while the proxy reaches the site. After that it runs the TLS handshake with the site through the tunnel and only then sends the request. In total that is four round trips to the proxy and three to the site before the first byte.

Each step is standardized. CONNECT is described in section 9.3.6 of RFC 9110, the SOCKS5 handshake in RFC 1928 with username and password authentication in RFC 1929, and the single round trip TLS 1.3 handshake in RFC 8446. SOCKS5 with a password costs two extra round trips to the proxy, for the greeting and the authentication. On a proxy 90 ms away an HTTP tunnel therefore opens about 180 ms sooner than a SOCKS5 one, and once the tunnel is open the two protocols are equally fast.

Two conclusions follow. Keep connections alive when your software allows it, because a scraper that opens a fresh tunnel for every URL pays the full setup each time, and a 10 ms proxy then behaves like a 40 ms one. A curl run without a session shows that setup cost; a browser with keep-alive shows the steady state.

Why might the ping be different?

The question is often asked: "Why does the ping jump?". For example, at the time of the first connection was 20 ms, and an hour later it is already 50. The ping changes because one leg of the route changed, and a proxy has more legs than a direct connection. Usually it is a rotated exit address, peak load on the proxy's network or a slower target site.

IP address rotation. If you use mobile proxies with automatic IP change every 5 to 10 minutes, the new address may be located in another part of the operator's network, which affects the delay.

ISP network load. For example, during rush hour, mobile networks experience increased loads, which affects data transfer speeds.

Response speed of the target site. High ping is not always related to proxies. If the site you are connecting to is slow, the resulting latency will increase, even if the proxies show minimal values. Chrome's DevTools documentation defines the waiting phase as one round trip plus the server's preparation time, so a slow site inflates every measurement through any proxy.

A fourth reason appears only when a direct connection is compared with a proxied one. Through a classic HTTP or SOCKS proxy the browser cannot use HTTP/3, which 40 percent of websites offer in 2026 according to W3Techs. It falls back to HTTP/2 over TCP, and part of the difference is the protocol rather than the proxy.

What is a good ping for a proxy?

A good ping for a proxy is one that stays under the limit of your task on the 95th percentile, not only on average. Under 50 ms suits trading bots, captcha work and games. A range of 100 to 150 ms is comfortable for scraping, account management and monitoring, and anything above 300 ms is felt as a visible pause in a browser and shows up as timeouts in automation.

Those limits have a physical side. Nothing travels faster than light in fiber, so a proxy on another continent cannot show 20 ms however good its server is, and the backbone values from the location section set the minimum for each route. Stuart Cheshire, the Apple engineer behind Bonjour, wrote in his 1996 essay It's the Latency, Stupid that "once you have bad latency you're stuck with it". Thirty years later the rule still holds for proxies, because you can buy more bandwidth but not fewer milliseconds on a long route.

On our proxies, ping is measured in the range of 5 to 300 ms depending on the type of server and task. If minimal latency is important, in-country server proxies are the best choice. Measure them from your own location with the curl loop above before buying a large package, and keep the server whose worst result, not its best, fits your limit.