Elasticsearch: Accelerating the Django Admin
summary
Django Admin search migrates from multi‑second queries to near‑instant lookups by sending terms to Elasticsearch, returning ids, and rehydrating via the ORM; production uses elasticsearch‑py not the DSL, versioned indices, and batch updates with Celery beat every two minutes; a problematic migration hit 12 second searches then fell to negligible latency; mappings get dropped and reimported as needed, 1M rows compress to under 100 MB in Elasticsearch.
AI-written summary · about
chapters (26)
- 0:00Introduction
- 0:44About Growth Collaborative
- 1:40The Django Admin
- 2:54Data Model
- 3:09Django Search
- 4:22Elasticsearch
- 5:11Terminology
- 6:44Document ID
- 7:14Search
- 8:34Get Search Results
- 9:00Updating Elasticsearch
- 12:24Customer Data Model
- 13:03Queries
- 13:20Elasticsearch DSL
- 15:04Elasticsearch Query
- 15:20Elasticsearch Website
- 15:36Takeaway
- 16:59Elasticsearch Versions
- 18:06Testing
- 19:58Hosting Options
- 23:34Resources
- 23:55Django Projects
- 25:13Switching from DSL to Pi
- 26:33Data Migrations
- 30:00Postgres
- 31:00Limits
related talks
42mElasticsearch DSL
29mMigrating Data Quickly and Safely
35mBeyond the ORM: from Postgres to OpenSearch with Andrew Mshar
39mSPONSOR WORKSHOP / Seth Larson / Elastic
48mEvolving Django: What We Learned by Integrating MongoDB
31mSpace Django: Migrating and Redesigning a Database while Hunting for Satellites.
26mDjango at
42mA Nice Problem to Have: Django Under Heavy Load