Beyond the basics with Elasticsearch
summary
Shows concrete recipes: index queries with percolator reverse search to drive alerts, live search and classification (geo by percolating a geo point against indexed city shapes; language by matching ≥4 language-specific tokens). Uses the Gauss decay function for distance scoring plus a small random boost; implements recommendations by combining significant_terms + sampler (top ~500 per shard) to avoid popularity bias. Notes memory pitfall for wide aggregations and fixes it by limiting a dimension first (breadth‑first sampling).
AI-written summary · about
chapters (18)
- 0:00Intro
- 5:32Relevance
- 13:26Percolator
- 13:51Alerting
- 15:40Geolocation
- 17:11Language Classification
- 18:12Event Classification
- 19:35Metadata
- 20:33Aggregations
- 22:06Elastic Surge
- 23:39Aggregation
- 31:09Graph Algorithm
- 32:33Aggregation and relevancy
- 34:18Question
- 34:37Trending
- 35:34Buckets
- 37:08Broader search
- 37:44Recap
related talks
37mBeyond the basics with Elasticsearch
01h 00mElasticsearch: you know, for more than search
35mElasticsearch - Beyond Full-text Search
36mRecommendations at scale with Elastic Search
47mAdvanced (Elastic) Search for Your Legacy Application
30mCaching for Speed: Practical Performance Optimization in OpenSearch & Lucene
54mAdvanced (elastic)search for your legacy application
51mAdvanced search for your legacy application