VPS DDoS protection explained
Every modern VPS provider advertises "DDoS protection." The numbers — 1 Tbps, 3 Tbps, 10 Tbps — look impressive. The reality is more nuanced. Here's what those numbers mean, what they don't, and what protection a typical self-hosted site actually needs.
Two completely different attack classes
Layer 3/4 — volumetric
Flood the link with raw packets. SYN floods, UDP amplification, NTP/DNS reflection. The goal is to saturate your network port or the provider's upstream. The defense is "more capacity upstream" plus on-net scrubbing that drops malicious traffic before it reaches you. This is what "3 Tbps protection" refers to.
Modern self-hosted sites essentially never see a real volumetric attack big enough to matter. Even a 10 Gbps flood requires real botnet resources; targeted at a $4 personal VPS, it's overkill the attacker would deploy against a corporate target.
Layer 7 — application
Send valid HTTPS requests at a rate your app can't handle. A few hundred requests per second to an expensive endpoint (search, login, AI inference) can take down a small site without filling any pipe.
This is what most self-hosted sites actually face when they face anything. Volumetric protection doesn't help here — the requests look like normal users at the network level.
What "3 Tbps" actually means
When a provider says "3.5 Tbps DDoS protection," it usually means: their upstream network capacity, across all their datacenters, can absorb up to 3.5 Tbps of inbound attack traffic without their network going down. Your individual VPS still has a 1-10 Gbps port; the protection lives upstream.
So "3 Tbps" doesn't mean your server handles 3 Tbps of legitimate traffic. It means an attacker can't take you offline by filling the wire upstream.
What you actually need
Self-hosted personal tools (n8n, Vaultwarden, WireGuard)
Basic volumetric protection at the provider level. Almost every reputable VPS host meets this bar; the cheapest plans often include it. Below 1 Tbps is fine; you'll never see an attack approaching that.
Small public site, blog, status page
Same as above plus a CDN in front (Cloudflare's free tier is the standard) to absorb traffic spikes and handle the L7 patterns that target small sites.
Telegram or Discord bot
You're outbound-only. No DDoS surface beyond SSH. Skip the worry.
Public API / service users depend on
Provider-level volumetric + CDN with rate limiting + application-level rate limiting in Nginx/Traefik. Three layers, each catching different things.
You attract specific attention (controversial topic, public-facing forum, etc.)
Real DDoS protection becomes meaningful. Look at Path.net, Cosmic Guard, OVH Game/Anti-DDoS, or Cloudflare Pro/Business. These are different from VPS-included "DDoS protection" in scope and cost.
What you can do at the application layer
Rate limit in Nginx
limit_req_zone $binary_remote_addr zone=app:10m rate=10r/s;
server {
location / {
limit_req zone=app burst=20 nodelay;
# ... proxy_pass ...
}
}
10 requests/second sustained, 20-request bursts. Tune to your traffic.
Rate limit by URL prefix
limit_req_zone $binary_remote_addr zone=login:10m rate=5r/m;
location /login {
limit_req zone=login burst=2 nodelay;
# ... proxy_pass ...
}
Strict on expensive endpoints (login, search), loose on cheap ones (static assets).
Cap concurrent connections
limit_conn_zone $binary_remote_addr zone=conn:10m;
server {
limit_conn conn 20;
}
Geofence at the edge
If your users are in the EU only, drop everything else. Use Cloudflare's country block rules or iptables + xt_geoip.
Cloudflare in front of a VPS
The standard "I have a small site and want defense in depth" pattern:
- Add your domain to Cloudflare's free tier.
- Cloudflare DNS proxies your A record; visitors hit Cloudflare's IP, Cloudflare forwards to your VPS.
- UFW on your VPS: allow 80/443 only from Cloudflare's IP ranges (refresh periodically from
cloudflare.com/ips). - Enable Cloudflare's "Under Attack" mode if you're actively being hit.
This setup hides your VPS IP from public DNS (less direct-to-origin attacks), absorbs L7 patterns Cloudflare recognizes, and gives you basic WAF rules — all on the free tier.
What VPS providers advertise vs what you should care about
| Marketing claim | Worth caring about? |
|---|---|
| "DDoS protection included" | Yes — baseline scrubbing is genuinely useful and often missing on the cheapest hosts. |
| "X Tbps DDoS protection" | Above ~500 Gbps the headline number stops mattering for any self-hosted use case. |
| "Layer 7 protection at the provider" | Rare and often expensive add-on; usually better handled by Cloudflare or in your app. |
| "Always-on scrubbing" | Yes — meaningfully better than "we route through scrubbing only when an attack is detected" (which has detection delays). |
| "Anycast network" | Yes — disperses attack traffic across many points of presence; harder to take down a single PoP. |
What's not DDoS protection
- fail2ban — protects you from credential brute-force, not from a flood of valid-looking requests.
- UFW — drops unauthorized traffic, but if 100k IPs each send one valid request, every one passes.
- A bigger VPS — scales linearly with attack budget; doesn't help against asymmetric attacks (Slowloris, etc.).
Layered defense is the answer: provider network + CDN + rate limiting in app + lean infrastructure. No single layer is enough.
VPS with included DDoS protection?
Inferno includes up to 3.5 Tbps volumetric protection on every plan, including the $3.49 starter. Pair with Cloudflare in front for L7 and you've covered both attack classes affordably.
What to do next
- UFW — the firewall layer.
- fail2ban — credential brute-force defense.
- Docker hardening.
- Best DDoS-protected VPS providers.
Author: Riku Lehtinen — Tampere. Last reviewed: 2026-05-22 against current major-provider attack reports.