What do you use to measure the health of your network both internally and externally? I put emphasis on health because I think it is a bit more complex than bandwidth, latency is also important as is jitter.
I used to use iPerf when I had separate and distinct devices for my workstation and router, but now they're a single box. I would periodically measure bandwidth nightly when the devices weren't being actively used and raise an email alert. My motivation for this back in the day was that I pulled the CAT6 cabling in my house and wanted to know if a cable or connector I did went bad. It's been well over 10 years and I think this is unnecessary.
For external bandwidth, I used to use fast-cli, but the current version presently requires go 1.22, so that isn't working for me at present. For external bandwidth, I would measure that hourly. I previously used a collection of pings to various websites to measure latency to determine if there was an issue.
With that said, in the past 10 years of this monitoring, ~ 4 on FreeBSD, the only issues I encountered were, 1, a former ISP's instability resulted in poor performance and 2, a USB NIC I used died. In the case where the NIC died, monitoring bandwidth and latency wasn't very helpful. I had to isolate the device to pinpoint the problem. For the former, the ISP wasn't very helpful, so capturing information didn't bear any fruit.
Coming back to my original question, I am inclined to drop monitoring and keep my systems simpler. It seems like unnecessary complexity for a marginal benefit.
I used to use iPerf when I had separate and distinct devices for my workstation and router, but now they're a single box. I would periodically measure bandwidth nightly when the devices weren't being actively used and raise an email alert. My motivation for this back in the day was that I pulled the CAT6 cabling in my house and wanted to know if a cable or connector I did went bad. It's been well over 10 years and I think this is unnecessary.
For external bandwidth, I used to use fast-cli, but the current version presently requires go 1.22, so that isn't working for me at present. For external bandwidth, I would measure that hourly. I previously used a collection of pings to various websites to measure latency to determine if there was an issue.
With that said, in the past 10 years of this monitoring, ~ 4 on FreeBSD, the only issues I encountered were, 1, a former ISP's instability resulted in poor performance and 2, a USB NIC I used died. In the case where the NIC died, monitoring bandwidth and latency wasn't very helpful. I had to isolate the device to pinpoint the problem. For the former, the ISP wasn't very helpful, so capturing information didn't bear any fruit.
Coming back to my original question, I am inclined to drop monitoring and keep my systems simpler. It seems like unnecessary complexity for a marginal benefit.