Top 4 ways to run a global latency test
Imagine you run a website or API, and users from another country report performance issues. Your local ping results may look fine, but they only show how your network reaches your API, not how someone in Japan, Brazil, or Germany does.
That’s why global latency testing is important. You need to run tests from different locations to really understand how your services perform across the world and not just on your machine.
In this blog post, we’ll explore four ways to do exactly that, from manual workarounds to dedicated tools, so you can pick the one that best fits your needs and workflow.
Option 1: Rent a VM and SSH in
How it works
Maybe the first workaround that comes to mind is to set up a cloud instance in the region you want to test from. AWS, Google Cloud, and DigitalOcean all let you do that fairly easily. Then you SSH into it and run ping from there.
When it’s useful
If you’re already working in a cloud console, this can feel like a natural thing to do. You already have access to the infrastructure, you know how to use it, and it usually lets you run ping and other commands.
Limitations
The downside of that approach is that it’s manual and slow. Setting up a cloud instance takes time, and it still limits you to testing from one location at a time. For example, to test from five regions at the same time, you’d need five instances. Also, it's likely that your cloud provider doesn't cover specific locations, especially if you want to target smaller regions. Moreover, it costs money, and while it is not much, it adds up, especially if you forget to tear an instance down after using it. This method may be okay for a one-off test, but it isn’t suitable for regular testing.
Option 2: Use a looking glass server
How it works
Many ISPs and network providers host publicly available looking glass servers that run at their edge, which allow you to run basic network commands from their infrastructure. You visit their page, fill out a form where you typically define the location, test type, and test target, and get results from that provider’s network.
When it’s useful
Looking glass servers are helpful when you need to check whether your traffic is actually reaching a destination through a specific provider's network and how it's being routed to get there. This can help you investigate potential routing issues with a particular ISP.
Limitations
Unfortunately, using looking glass servers can be a frustrating experience. Every provider has a different website, UI, and set of supported commands. If you want to test across multiple networks, you have to search each provider’s tool separately and run your tests on each of them. All in all, this approach isn’t very consistent, and results may not always be easy to share or compare.
If you’re looking for a more streamlined experience, you can find looking glass pages for hundreds of network providers on the Globalping website. There, you’ll find a consistent user interface and uniform test result output. You can run all Globalping test types, like ping, traceroute, MTR, DNS, and more. It won't replace the full power of the Globalping CLI (more on that below), but for provider-specific testing, it's a much more convenient starting point.
Option 3: Test on the Globalping website
How it works
Visit globalping.io and fill out the form at the top of the page to run a ping test against your target. Of course, you can run other test types as well, define different locations and multiple ones, and define the number of probes to run tests on. Results will appear within seconds, showing the used probes and test results on a map, with the individual test results listed below.
When it’s useful
Using the website tool is the fastest way to get a global picture of your services’ latency without installing anything. You can test from multiple locations at once and view results on a map or in a list. Each measurement also includes a link you can share with others or revisit later.
Globalping is free to use and requires no account, making it a great option for quick checks and for users who prefer working in a UI.
Another benefit worth mentioning: unlike a standard ping command, Globalping lets you run ping with either ICMP (the default) or TCP. This is handy if ICMP ping is blocked by a firewall, for example. You can read more about when and why to use TCP ping in our dedicated post.
Limitations
If you need to run frequent tests, filling out the website form each time can be time-consuming and tedious. In that case, the Globalping CLI is a better fit.
Option 4: Run tests using the Globalping CLI tool
How it works
Install the Globalping CLI on your machine, and you can run global ping tests directly from your terminal. For example, a basic command looks like this:
globalping ping globalping.io --from Europe --limit 2
This will run ping to the Globalping website from two random probes in Europe. You can also fine-tune location selection and add other options to customize your commands.
To learn more, simply run globalping help for general information on available commands, and globalping ping --help to learn more about a specific command like ping.
When it’s useful
If you spend a lot of time in the terminal anyway, this option is probably the best fit for you. You can script tests, get results output as JSON for easy parsing, and integrate them into your existing workflows or CI/CD pipelines. You get all the same benefits as the website tool, including free, global coverage, multi-location testing, and TCP ping support, all without ever leaving your terminal.
Limitations
If you’re running a lot of tests in a short time, you’ll eventually hit the free hourly test limit (currently at 250 tests per hour). However, you can increase this allowance by signing up for the Globalping Dashboard for free and connecting it to your Globalping CLI (which currently gives you 500 tests per hour). If that isn’t enough, you can host a probe to support the Globalping network and earn tokens for additional tests.
Conclusion
If you want to run a global latency test with ping, you have a couple of options:
- SSHing into a rented VM: works, but doesn't scale and can get costly
- Looking glass servers: good for provider-specific checks, but fragmented and inconsistent
- Globalping website: instant global results with no setup, but can be tedious for frequent testing
- Globalping CLI: instant global results, scriptable, and ideal for terminal-native workflows
No matter which method you use, the main point is the same: testing with only local ping often isn’t enough. Testing from multiple locations gives you a clearer picture, and tools like Globalping offer a couple of options to do so.
Ready to give it a try? Run your first global ping test for free on globalping.io.