DEV Community

Arina Cholee
Arina Cholee

Posted on

How We Stopped a Flood of Bot Traffic Using SafeLine

We run a small web platform. Nothing crazy — just a login page, some APIs, and a dashboard. Like a lot of small teams, we don’t have a dedicated security engineer. It’s mostly one backend dev who “owns security because he knows Linux better than everyone else.”

For months, we were getting hammered by:

  • Bot-driven brute-force attempts
  • HTTP flood spikes that slowed the site to a crawl
  • Random scanners probing for SQLi/XSS
  • Rising bandwidth bills for no good reason
  • A cloud WAF that threw false positives at legit users

Every week looked like:
“Why is the login page on fire again?”

Eventually we got tired of fighting the same battles and started looking for something open-source, self-hosted, and not annoying.

That’s how we ended up trying SafeLine.

Why We Tried SafeLine

We had three basic requirements:

  • Self-hosted → full control, no opaque vendor rules.
  • Strong anti-bot and anti-flood protection → more than IP blocklists.
  • Easy deployment → no multi-week project.

SafeLine checked all three, so we set up a test.

Deployment took minutes.
We routed traffic through it…
And waited.

Within the first hour, logs made one thing clear:

Most of our traffic wasn’t human.

What Actually Helped

1. Legitimately strong semantic detection

SafeLine doesn’t rely on heavy signature sets. It analyzes request semantics, and it caught weird payloads our previous WAF missed, including:

  • Obfuscated SQL payloads
  • Parameter smuggling
  • Odd XSS vectors
  • Scanner noise we stopped noticing long ago

False positives dropped close to zero.

2. Rate limiting that works in the real world

We set rules like:

  • 5 req/s → login
  • 20 req/s → API
  • global per-IP thresholds

Bots hit the limits instantly.
Legit users didn’t feel a thing.

During unexpected traffic spikes, the Waiting Room absorbed the surge instead of dragging the whole site down.

3. Bot challenge that actually stops bots

We enabled SafeLine’s bot challenge on our login page.

Bots: stuck.
Users: completely unaffected.

This alone killed ~70% of our daily noise.

The Results After 30 Days

Here’s what changed:

  • 90%+ malicious traffic blocked
  • Zero downtime from traffic floods
  • Lower bandwidth usage since junk traffic never hit origin
  • Ops time saved — fewer “urgent investigation” alerts

Biggest win:
Our team finally got to stop firefighting and focus on product work.

Why We’re Staying With SafeLine

A few reasons:

  • Fully self-hosted
  • No vendor lock-in
  • Predictable cost (no per-domain billing)
  • Works cleanly with existing reverse proxies
  • Free version already covers core WAF features
  • Pro version available if needed

It feels like something built by people who actually run servers under real attack.

What’s Coming Next

The SafeLine team mentioned future work like:

  • Identity & Access Management (IAM)
  • A more complete Web Security Gateway approach
  • Improved bot behavioral detection
  • More visualization/monitoring tools

If they deliver even half of that, it will replace several tools we currently use.


Try It Yourself

SafeLine is open-source and free.
You can find it on GitHub (search for: SafeLine WAF).

If your stack includes:

  • a website
  • an API
  • a dashboard
  • anything bots love

…it’s worth running a quick test.

Top comments (0)