Winter Mode: Running on Just One Server
In winter the roof delivers hardly any power, so three servers have to make do with the work of one. What that looks like – and everything you can see in the graphs along the way.

🌐 Also available in: Deutsch · Français · Español · Italiano · Português · Nederlands · Polski
In short: In winter my PV system delivers little power, so only one of my three servers runs and carries the consolidated workloads. Grafana nicely shows the Nextclouds’ backup cycles and the nightly virus scan. Plus a short guide on how to set up Node Exporter, Prometheus and Grafana in a few minutes.
Right now I’m only running one of my three servers…
thanks to winter season my PV setup isn’t exactly flooding me with power,
so I’ve had to consolidate my workloads a bit. (Kidding… mostly.) ☃️🔋
But it is pretty funny watching the whole thing in Grafana,
which I use to monitor the server’s performance.
From the network load you can clearly see the 6-hour backup cycles of the two Nextcloud instances I manage for local clubs.
The high CPU load during the night — up until shortly after 2 a.m. — was simply the antivirus scanner doing its job.
The second big spike around 1 p.m. was my lunch break, when I watched a movie via Jellyfin in the browser.
The browser forces heavy transcoding, meaning the server has to adjust the video quality for the client — especially when subtitles are being “burned in,” which costs quite a bit of CPU.
Now that I’ve installed the Jellyfin Media Player, none of that CPU load happens anymore — the server stays completely relaxed. 🎥⚙️😄
To get proper observability, I also installed Node Exporter, Prometheus and Grafana on the server.
The whole setup took only a few minutes — seriously! ⚡🛠️
Here are the key steps I followed:
- Installed Node Exporter to expose hardware and system metrics.
- Set up Prometheus and added Node Exporter as a scrape target.
- Installed Grafana and connected it to Prometheus as a data source.
- Imported a few prebuilt dashboards to visualise CPU, RAM, disks and network usage.
With these three components running, I have a clean, real-time view of what the machine is doing —
perfect for watching all those tiny spikes and dips that make a sysadmin’s heart happy. ❤️🖥️📊
At the moment I can’t push the system any harder, even though it’s running:
- Jitsi
- GitLab
- my NAS
- a fresh Icinga2 base installation
- Pi-hole
The machine still has a surprising amount of resources left in the tank… 🚀💪
🎯 Conclusion
One server instead of three is plenty in winter – and with Node Exporter, Prometheus and Grafana you can see exactly what it gets up to at night. The monitoring was set up in a few minutes and turns guesses about load and backups into clear curves. ☃️
📚 Related posts:
