2.9K subscribers • 3 videos on Hallway
Intermediate. Assumes working knowledge.
14K views10 years ago4 recommendations
Databases and dataDistributed systemsPostgresDatabase internalsScaling and large-scale architecture
Dan Robinson explains how Heap scales its product analytics backend to billions of events using PostgreSQL, Citus, and Kafka.
Powering Heap With PostgreSQL And CitusDB At Heap, we lean on PostgreSQL for all our backend heavy lifting. We support an expressive set of queries — conversion funnels with filtering and grouping, retention analysis, and behavioral cohorting to name a few — across billions of users and tens of billions of events. Results need to come back in a matter of seconds and reflect up-to-the-minute data. This talk will discuss these challenges, with a particular focus on: - Using CitusDB for interactive analysis across 50 terabytes of data and counting. - PostgreSQL and Kafka: two great tastes that taste great together. - UDFs in C and PL/pgSQL, partial indexes for pre-aggregation, and other tricks up our sleeves.What people say. Recommended as a great case study on scaling event processing and analytics using PostgreSQL and Citus.
Judsonon Hacker News • Recommends
On “TrailDB – An Efficient Library for Storing and Processing Event Data”
devbugon Hacker News • Recommends
On “Building a Scalable Postgres Metrics Backend Using the Citus Extension”
simianoon Hacker News • Recommends
On “When to Avoid JSONB in a PostgreSQL Schema”
macoboon Hacker News • Recommends
On “How Discord Stores Billions of Messages Using Cassandra”
macoboon Hacker News • Recommends
On “Running 10M PostgreSQL Indexes in Production”
malisperon Hacker News • Mentions
On “When to Avoid JSONB in a PostgreSQL Schema”
malisperon Hacker News • Mentions
On “When to Avoid JSONB in a PostgreSQL Schema”
malisperon Hacker News • Mentions
On “Launch HN: PostHog (YC W20) – open-source product analytics”