I'm currently managing the IT for a small to medium enterprise (SME) with around 100 users in a somewhat compact office. We have a basic network setup with a firewall up front and three access points to serve everyone. There's no on-premise infrastructure at all.
I've started implementing internet quality monitoring, but I'm wondering if I'm on the right track or if there's something more efficient out there. So far, I've created a few scripts:
- One runs every 15 minutes to carry out speed tests using the Speedtest CLI, though I know this may not be the most accurate gauge of quality.
- Another runs every 5 minutes to ping a handful of popular websites, logging the average response time.
- A third script executes web service requests every 5 minutes against common sites to check their load times.
I have alerts set up that notify us via email if any script reports unusually high ping times or if a site takes too long to load. All the data is displayed on a dashboard, allowing us to track internet quality over time in terms of speed, pings, and web request performance.
Another team is working on deploying PRTG in about a month, but I'm curious: what are some other ways to monitor internet traffic passively? Am I unnecessarily complicating things by not using readily available freeware?
4 Answers
Before diving deeper, you might want to consider why monitoring is essential for your setup. If you're getting complaints about internet issues and call quality, you can complement your web service metrics with this. It's a good idea to differentiate between whether the issues stem from the internet, the endpoints, or specific services.
Have you considered using more advanced tools like LibreNMS or Prometheus? I found they work well for larger setups like yours, though you might need a dedicated machine to run them.
Be cautious about running speed tests frequently, as they could potentially slow down the connection while testing. However, I've used speedtest-cli and it usually consumes minimal data, so it's not too bad. Just something to keep in mind!
That’s true! But from my experience, it's generally okay without overly bogging down the network.
You might also want to set up your own internal speed test server. This way, when users complain about slow internet, they can run a test directly against it to see if the problem lies within your network instead of the outside internet. This could be a useful diagnostic tool, but starting with the internet connection makes sense too.
That sounds useful! I’d imagine it could help eliminate any internal network issues before delving deeper.
Exactly! We've managed to pinpoint issues this way too, using app-specific metrics, like Zoom's call logs for instance.