Scale Elasticsearch for Your Java Applications. Philipp Krenn. Java Fest 2020 Online

on YouTube ↗54m

" summary

Shows a small Java injector that uses the Elasticsearch high‑level REST client and proves concrete wins: switching from single-document HTTP calls to bulk requests cuts 1,000 inserts from ~49s to ~11s (and can reach ~4s with tuning). Demonstrates changing the refresh interval for large imports, why default sharding moved to one primary+one replica, shard sizing guidance (avoid tiny shards, aim >10–50GB, limit shards per GB of heap), and replacing deep pagination with search_after or scroll. Uses Kibana monitoring and Rally for benchmarks.

AI-written summary · about

" chapters (15)

1. 0:00Intro
2. 0:12for Your Java Applications
3. 1:38Terminology Cluster, Node, Index, Shard
4. 5:02Disclaimer This is not a benchmark
5. 13:40Stack Monitoring
6. 20:43Optimal Bulk Size?
7. 22:37index.refresh_interval Default is
8. 28:08Load Distribution
9. 30:27Topology Sniffing Cloud Proxy
10. 31:18Distribution Round Robin Adaptive Replica Selection for searches
11. 33:34Index Patterns Static vs Time-Based
12. 39:23Pagination
13. 42:37For Your Protection index.max_result_window: 10000
14. 47:23Rally Macrobenchmarking Framework for Elasticsearch
15. 48:02Nightly Benchmarks

" related talks

• 54mAdvanced (elastic)search for your legacy applicationAravind Putrevu · Voxxed Days Singapore 2018
• 47mAdvanced (Elastic) Search for Your Legacy ApplicationDavid Pilato · DevDays Europe 2022
• 01h 01mBest Practices With ElasticsearchRoy Russo · Devnexus 2020
• 16mSo many results... Give them to me!Daniël Spee · J-Fall 2022
• 29mHow (not) to Scale Elasticsearch for Data Analytics!Aishwarya Sankaravadivel · Developer Summit 2020
• 51mAdvanced search for your legacy applicationDavid Pilato · Jfokus 2018
• 35mElasticsearch - Beyond Full-text SearchAlex Reelsen · GOTO Aarhus 2013
• 01h 00mAdvanced search for your legacy applicationDavid Pilato · Joker 2016