[ 03 // HARD METRICS CASE STUDY ] · CLIENT: Sestry.eu (Netguru Contract) · ROLE: Edge Performance & Infrastructure Technologist · 2024 – 2025

Sestry.eu: Engineering High-Traffic Edge Infrastructure, AI Bot Shields & Onet.pl Syndication

A multidisciplinary engineering approach to international media syndication, edge caching, and AI scraper defense.

CONTENT SYNDICATION Real-time ATOM Feed to Onet.pl

BOT DEFENSE Custom WAF blocking AI LLM crawlers

LOCALIZATION DeepL API translation pipeline

CACHE EFFICIENCY Sub-50ms Edge Response across EU

[ 01. THE BUSINESS STAKE ]

The Commercial Challenge

Sestry.eu is an international media initiative delivering critical news and journalism across Poland and Ukraine. Reaching millions of readers required seamless distribution partnerships with tier-1 media portals (specifically Onet.pl / Ringier Axel Springer), strict uptime under high concurrent traffic spikes, and immediate protection against rogue AI web scrapers draining server bandwidth.

[ 02. SYSTEM ARCHITECTURE ]

Data Flow & Infrastructure Blueprint

[ Global Readers / High Traffic Spike ]
               │
               ▼
       [ Cloudflare Edge Network ]
        ├── [ WAF & AI Bot Shield ] ──▶ (Drop unauthorized LLM scrapers)
        ├── [ Cache-Everything Edge Rules ] ──▶ (Sub-50ms cache hits)
        │
        ▼ (Origin Proxy)
   [ Webflow CMS / Media Origin ]
        │
        ├──▶ [ Cloudflare Worker: ATOM XML Transformer ] ──▶ [ Onet.pl Automated Syndication ]
        └──▶ [ DeepL API Localization Engine ] ──────────────▶ [ Real-time Polish/Ukrainian Sync ]

[ 03. THE ENGINEERING HURDLE ]

Breaking Points & Architectural Solutions

Onet.pl required strict compliance with a custom ATOM/RSS XML schema that standard Webflow RSS feeds could not generate. Additionally, massive spikes during breaking news events threatened to exhaust origin bandwidth limits, while unauthorized automated bots repeatedly attempted to scrape content for LLM training. I engineered a Cloudflare Worker that dynamically fetches raw CMS content, converts the payload into Onet's exact proprietary ATOM XML specification on the edge, and caches the result with atomic invalidation. In parallel, I implemented custom Cloudflare WAF firewall rules to fingerprint and block automated AI scrapers, safeguarding proprietary editorial content.

[ 04. QUANTIFIED BUSINESS OUTCOMES ]

Measurable Impact

▶ Opened a multi-million reader distribution channel via seamless automated syndication to Onet.pl.

▶ Protected origin servers from downtime and high overage costs through aggressive Cloudflare Worker edge caching.

▶ Automated editorial translation workflows with DeepL API, drastically cutting manual editorial overhead.

[ 05. TECHNOLOGIES & TOOLS ]

Cloudflare Workers Cloudflare WAF Edge Cache API XML / ATOM Syndication DeepL API Reverse Proxy

[ NEED THIS ARCHITECTURAL RIGOR? ]

Book an Architecture Discovery Call

Direct review of your AI platform, design system, or edge infrastructure.

Book Discovery (Cal.com)