Tagged articles
search performance
10 articles · Page 1 of 1
A large‑scale e‑commerce search system handling 3 billion product documents and 500 million daily queries was on the brink of collapse, but a systematic overhaul across query modeling, index design, write pipelines, cluster topology, and application‑level governance lifted throughput five‑fold while cutting P99 latency from 3.2 seconds to 180 ms.
ElasticsearchJavabulk processing
0 likes · 30 min read

The blog benchmarks Quickwit by indexing 61 billion GitHub event documents (23 TB) with 27 MB/s throughput, achieving sub‑second search latency using eight CPUs at $0.0002 per TB per query, while indexing costs about $2 per TB and storage costs $8.4 per TB per month.
GitHub ArchiveQuickwitbig data
0 likes · 19 min read

This article details how upgrading from Elasticsearch 5.x/2.x to 8.x dramatically improves search, aggregation, and write performance while reducing storage costs, introduces vector KNN, synthetic _source, TSDS, searchable snapshots, security enhancements, and provides migration examples and code snippets for enterprise search platforms.
ElasticsearchHybrid RankingSecurity
0 likes · 18 min read

By defining index sorting on the publish_time field when creating the Elasticsearch index, the team transformed a multi‑second full‑scan query into a sub‑50 ms operation, demonstrating that pre‑ordered storage dramatically speeds up large‑result‑set sorts while modestly affecting write throughput.
DocValuesElasticsearchIndex Sorting
0 likes · 12 min read

iOS WeChat’s full‑text search was upgraded by selecting SQLite FTS5, creating a VerbatimTokenizer with multi‑level delimiter support, optimizing table formats to cut index size by 30 %, improving batch index updates and parallel search logic, resulting in 40‑60 % faster query latency.
Index OptimizationSQLite FTS5WeChat
0 likes · 26 min read

This guide explains how to allocate hardware resources, calculate Elasticsearch cluster size based on data volume, and conduct indexing and search benchmarks using Rally to ensure production‑grade performance and capacity planning.
Cluster SizingElasticsearchRally
0 likes · 12 min read


政采云技术Jul 14, 2022 · Operations
Facing severe IO-wait and read bottlenecks as product data grew from tens of millions to billions, this article analyzes root causes in Elasticsearch clusters and presents a comprehensive solution involving index parameter tuning, merge settings, translog async writes, query optimizations, and hardware upgrades to restore performance and stability.
Cluster OperationsElasticsearchIO optimization
0 likes · 14 min read

This article describes how a top architect optimized Elasticsearch for handling billions of records, covering Lucene fundamentals, index and shard design, DocValues, query performance tuning, bulk indexing strategies, hardware considerations, and testing methods to achieve sub‑second query responses across multi‑year data ranges.
ElasticsearchIndex OptimizationLucene
0 likes · 12 min read

This article explains why deep pagination in Elasticsearch leads to high CPU usage, introduces core concepts such as indices, shards, nodes, and documents, and details the distributed CRUD operations, query and fetch phases, and segment management that affect search performance.
Deep PaginationElasticsearchIndexing
0 likes · 12 min read

This guide explains how to improve Elasticsearch indexing speed, search speed, and overall cluster performance by adjusting bulk request sizes, refresh intervals, replica settings, hardware resources, mapping choices, caching strategies, and query preferences, while also offering general best‑practice recommendations.
Elasticsearchhardware optimizationsearch performance
0 likes · 18 min read
