A slow VPS can mean your own software needs more resources, something misbehaves inside the server, or – less often – the host node is busy. A few measurements tell these apart quickly, so you fix the right thing instead of guessing.
Before you start
- SSH access and the Virtualizor panel; see resource monitoring.
- Note when the slowness happens: always, at certain times, or only for some pages.
Check the panel graphs
In Virtualizor, open the VPS Statistics. Look for CPU at its limit, RAM full, high disk I/O or network spikes at the times you noticed slowness. Correlating time is the fastest clue.
Measure inside the VPS
Run top (or htop) and read the header line:
| Value | Meaning when high |
|---|---|
us / sy | Your processes use the CPU – find them in the list |
wa (I/O wait) | Processes wait for disk – heavy database or backup activity |
st (steal) | The host is giving CPU time to other VMs – persistent values above ~10% are worth reporting |
| load average | Compare with the number of cores (nproc) |
Then check memory with free -h (heavy swap use means too little RAM) and disk space with df -h (a full disk slows or breaks databases).
Common causes and fixes
- A busy process: a PHP site under bot traffic, a stuck script or malware. Identify it in
top, check logs, block bad bots, or kill and fix the process. - Database load: slow queries or no caching. Tune MySQL/MariaDB; see tune MySQL memory.
- Not enough RAM: reduce services, add swap as a buffer, or upgrade; see upgrade your VPS.
- Disk full or I/O heavy: clean logs and old backups, schedule backups at night.
- Network: test with
mtrto your location; packet loss far from the server is outside the VPS. - Bandwidth throttled: check whether you reached the monthly limit in the panel.
When to contact support
If your own usage is low but you see high st steal time or high I/O wait without matching activity in your VPS, the host node may be overloaded. Open a ticket with screenshots of top, the times it happens and the VPS IP. MdmHost customers can use a support ticket.
Prevent future slowdowns
- Set up monitoring and alerts; see Grafana monitoring.
- Keep software updated and remove unused services.
- Size the plan for peak load, not average; see RAM requirements.