Scaling Grandma’s Hefezopf
summary
A product-widget integration for recipe sites scaled to handle a sudden 6 million views against a ~60 million‑product Solr index and 330k recipe URLs. It uses RAKE inside Solr for single‑digit‑millisecond keyword extraction and context‑enriched query boosting; the team replaced a manual URL‑mapping MVP, migrated the API to a partner LPA and added caching to stabilize throughput, raising conversion on the large partner above 2%.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 3:54Shopping24 tech stack
- 7:22Traffic surge at Chefkoch.de
- 10:31Agile scaling
- 13:43Manual mappings issues
- 15:54Recipe keyword extraction
- 20:30RAKE implementation
- 24:00New solution benefits
- 25:23Holiday traffic lessons
- 27:07Crawler and caching
- 28:46Widget conversion rate
- 30:11Summary
related talks
28mRe-examining recipe search
19mMigrating from legacy platforms to node
20mGain speed and space with NLP in Solr #bbuzz
31mScaling Facets to the Stars 🌟
30mMaking a Brand Programmable
45mBuilding amazing searches with Search API
23mRevolutionizing API Search with Holon Concept Explorations
40mBuilding a search platform 101 #bbuzz
