Loading page...Загружаем страницу...正在加载页面...
AtlasRepo

I gave Fail2Ban a memory: the story of ARE and the night I ended up banning myself

Description and use

ARE added memory of IP behavior, enabling more accurate threat assessment and reducing false positives, but during testing the author accidentally blocked his own IP.

Problem

Fail2Ban and other sensors lacked a memory of IP behavior, leading to isolated decisions that could misclassify legitimate traffic or miss persistent abuse.

Outcome

ARE successfully accumulated detection signals, applied decay, and escalated a self‑generated attack to a temporary ban, demonstrating the value of a memory‑based reputation layer.

Workflow

  • 1. Configure Cloudflare to allow own IP to reach origin.
  • 2. Generate non‑destructive attack traffic (SQLi, LFI, XSS, etc.).
  • 3. ModSecurity logs detections; Fail2Ban applies its rules.
  • 4. ARE collects sensor events, updates reputation scores per category.
  • 5. ARE evaluates reputation, applies decay, and transitions IP through decision states.
  • 6. Observe final state (e.g., BAN) and recover via out‑of‑band console if necessary.

Evidence

I didn’t want to add yet another security tool to the pile. I wanted to solve something that, from my point of view, was missing: that the server could remember an IP’s behavior and make decisions taking its history into account, not just the last event recorded.

Source

ARE was already running on my production VPS. So I decided to run a controlled test from an IP of my own, without involving any third parties and using only non‑destructive requests.

Source

The decision timeline reflected the progression: WATCH (18, LOW) ↓ FILTER (27, MEDIUM) ↓ BAN

Source

  • Record date: 2026-10-25T18:11:37.838Z