In this presentation, Meenakshi Jindal will discuss how Netflix scaled search queries with the growing scale of digital media assets using Elasticsearch. Asset Management Platform organizes and manages assets metadata, schema, relations, and permissions of images, videos, and text. Initially, we started with an index sharding strategy based on the asset types, but differences in the assets count by type created an unbalanced shard size on the cluster nodes. For queries across the asset types, we start noticing timeout and CPU spikes on the cluster. Meenakshi will share the journey of how they reevaluated their indexing strategy, and created indices based on time buckets instead of asset types, leading to better performance and searchability.
- Strategies for scaling Elasticsearch horizontally and vertically
- Designing indices, their settings, and mappings for efficient searchability
- How to balance shards size for better query performance
- How to handle indexing challenges when dealing with large data sets and multiple doc types.
- Framework to evolve with metadata search requirements e.g. updating some fields from `keyword` type to `text` type for full-text search requirement.
- 5:30 PM - Doors open
- 5:45 PM - Speaker Intro (Meenakshi Jindal)
- 5:50 PM - 6:30 PM - Presentation
- 6:30 PM - 7:30PM - Enjoy refreshments and mingle to get to know one another!
88 Kearny St · San Francisco, CA 94108 (Elastic Office - 19th Floor)
Directions & Parking:
This Elastic office does not offer parking. Below are a few recommendations for close public parking garages. Please keep in mind that parking is not allowed to be expensed and must be paid for on your own.
- The White House Garage: 223 Sutter St, San Francisco, CA 94108
- Post Montgomery Center Garage: 173 Sutter St, San Francisco, CA 94101
- Paramount Garage: 177 Jessie St, San Francisco, CA 94105
Bart & Muni Information:
- The nearest BART station is Montgomery St. Station
- The nearest Muni Station is Market & New Montgomery