Possible if you don't measure. My Contabo VPS had blips every few months, 30 seconds here, two minutes there. Their status page never showed yellow. I started logging externally, then I knew.
3 Years Stuck With Contabo – Here's What Actually Happened
External monitoring is the only monitoring. I use UptimeRobot and a self-hosted Uptime Kuma. If both agree it's up, it's up. If they disagree, I investigate.
GuruDeutschlag
Member OP
Ja, this is correct. I use HetrixTools and my own Nagios. Any host's status page is a marketing tool first, information source second. Never rely on it alone.
So what did we conclude about the OP's actual situation? Contabo VPS, moved to Hetzner, found out Hetzner's "dedicated" might also be virtual?
petegrid1990
Member
That's where it drifted, yes. The OP's three-year story became a general audit thread. Which is better than most forum drifts.
I'm the OP's cousin or something, I read the first post. They said Contabo "dedicated" was actually KVM, then moved to Hetzner and found similar. That's the pattern.
GuruDeutschlag
Member OP
Thank you for clarifying, maxkkin. This makes the thread more coherent. So both providers used "dedicated" language for virtualised products. The lesson is universal.
Universal lesson: if the price seems wrong for real hardware, it's not real hardware. Physics and accounting don't negotiate.
Watts are the real currency. A real EPYC at idle pulls more than some of these "dedicated" plans cost total. The math doesn't work.
one small ping for man...
pieter_rtm
Member
Rotterdam. I did the math once for a blog post. Real EPYC 7313, 64GB DDR4, 2x NVMe, 1Gbps unmetered in Falkenstein. Floor cost around 85 EUR just for power and space at retail rates. Anything under 120 EUR monthly is subsidised, used, or virtual.
Containers before it was cool