|Server Name||Provider||Status||Uptime||Load||NIC||RAM||Disk Space Free||IOWrite||IOPS||WAN||Downtime|
I run a wget to CacheFly's test file to test the WAN speed, a really quick dd to test the IO write speeds, and a quick ioping to test the IOPS every 2 hours (with a random delay thrown in for good measure). They aren't perfect benchmarks but, because all of the VPSs are tested somewhat equally (some providers use ploop on OpenVZ that results in incorrect IO results), I can at least get an idea of how they are performing. A blank spot means that the test didn't finish during the last scheduled time so it should update again within 2 hours.
Every minute a job is run on this server that uses 3 out of 11 servers randomly and checks the connectivity between those 3 servers and the target IP, if all 3 servers fail to connect to the IP then it adds one minute of downtime to the VPS. It's fairly accurate since the 11 servers are geographically different and it's checking in 1 minute intervals. Sometimes I reset the downtime manually if the "downtime" is a result of a mistake on my part (like forgetting to change a VPS's IP when a provider changes it, if the IP change is unplanned then I won't reset it though) but the downtime you see is total as of 07/29/2016, not weekly/monthly/yearly/etc...
The "Status" (i.e. does it ping?) is updated every minute. The rest are updated roughly every 5 minutes.
Sure, I'll remove it from this list but I have another list that only I can access so I can use that for my own use. If the scheduled performance tests are a concern I can always increase the scheduled time or adjust the tests.
I am a co-owner of Secure Dragon and no I will not mess with any results on this site because it doesn't benefit me if the VPSs I use have lower benchmarks. As you can also see 5 of the VPSs on this page are hosted with Secure Dragon and you can see all of our node stats here.
For the server communication I'm using an extremely hacked version of scrd (the one on GitHub hasn't been updated in 4 years and the developer stole a lot of money from us so I do not recommend using it). The website is some custom code I've been building for the past few years. I built a special system and daemon for the uptime monitoring.
There's some contact info on my website here: JMD.cc