Architecting real-time analytics for your customers
Time: 13:30 - 16:30
Feedback: Leave feedback
We'll cover various schema designs for ingesting large amounts of data (millions of records per second), how to perform rollups of data both in order and out of order, how to bucket rollups so they're composable, and finally some various sketch algorithms and datatypes that can be valuable for these type of applications such as TopN and HyperLogLog. We'll spend most of our time in SQL including scheduling our jobs to do all the heavy lifting directly in the database with pg_cron. Finally we'll look at how we can further scale this our and parallelize the workload for better performance with Citus.