Citus and JSON for real-time analytics at Vizor Games

on YouTube ↗31m

" summary

Vizor Games runs Citus on bare‑metal Gentoo with ~40 workers and 20+ tenant databases, ingesting about 100GB per day of semi‑structured JSONB and getting ~6:1 compression from Btrfs to control storage. JSONB causes planner blind spots so fixes include GIN jsonpath and per‑key Btree indexes, selective denormalization (largest table ~28 TB) and materialized‑view emulation via distributed INSERT…SELECT; the cluster delivers most analytics under 30s with scheduled ETL and custom maintenance.

AI-written summary · about

" chapters (10)

1. 0:00Introducing Ivan
2. 0:29Talk begin
3. 2:23Requirements for analytics @ Vizor Games
4. 6:09Details of real-time analytics solution
5. 8:19Performance Results
6. 8:46Special JSONB considerations
7. 15:08Maintenance: Roles, Tools, & multi-tenancy
8. 22:53User experience: Expectation vs Reality
9. 24:54Monitoring memory usage & query execution
10. 28:35Q&A

" related talks

• 55mBreakout SessionDan Robinson · PGConf Silicon Valley 2015
• 30mBuilding geospatial apps on Postgres, PostGIS, & Citus at large scaleNirav Savani, Mahipal Ramidi · Citus Con: An Event for Postgres 2022
• 25mData modeling, the secret sauce of building & managing a large scale data warehouseMin Wei · Citus Con: An Event for Postgres 2022
• 28mCitus Cluster: our zero-downtime migration storyMatthieu Dumont, Antoine Hemery · Citus Con: An Event for Postgres 2022
• 25mBig Spatiotemporal Data Management Using MobilityDB & CitusMohamed Bakli · Citus Con: An Event for Postgres 2022
• 24mScaling to Billions of Internet Observations per DayHudson Clark, Ethan Uberseder · Distributed SQL Summit 2020
• 55mThe JSON Evolution with SQL DatabaseRoberto Freato · Developers Summit 2021
• 27mFast DIY O11y with OpenSrc Analytics & VisualizationRobert Hodges · SREday San Francisco 2025 Q2