Performance & reliability
Is a VPS secure?
The isolation is strong: KVM enforces separation in hardware, and escaping it is a serious research-grade exploit. The realistic risk is not the hypervisor — it is your configuration. Password SSH, unpatched software and an open service you forgot about account for essentially all real-world compromises.
Assume the provider can technically read an unencrypted disk, because that is true of every virtualisation platform including every hyperscaler. Full-disk encryption with a passphrase the provider never holds is the only structural answer, and it only protects data at rest.
Everything else is ordinary server hygiene: keys not passwords, default-deny firewall, automatic security updates, minimal exposed surface, and tested backups. Those five things are worth more than any provider feature.
See also: Hypervisor · KVM
7 answers in Performance & reliability
- What is CPU steal time and why does it matter? Steal time is the percentage of time your virtual CPU was ready to run but the hypervisor gave the physical core to someone else. It appears as the "s…
- Why is my VPS slow? In order of likelihood: memory exhaustion causing swap thrashing, disk IO saturation, CPU steal time on shared cores, an unoptimised database, or netw…
- How much latency should I expect from a VPS? Within a metro, under 5 ms. Within a continent, 10–40 ms. Across the Atlantic, 70–90 ms. Europe to Asia, 150–200 ms. Antipodal routes, 250–320 ms. Lig…
- How many visitors can a VPS handle? A well-configured 2 vCPU / 4 GB instance serves a static site to millions of monthly visits, a cached WordPress site to roughly 300,000, and an uncach…
- How do I benchmark a VPS? Measure the three things that actually vary: single-thread CPU with sysbench, disk with fio at 4K random read/write, and network with iperf3 to a know…
- How much faster is NVMe than SATA SSD? Roughly five times the random IOPS — about 500,000 versus 90,000 — and around five times lower latency, under 0.1 ms versus about 0.5 ms. Sequential t…