How to build a social network on serverless
summary
Migrates a monolith to an event‑driven serverless backend using Kinesis streams, Lambda handlers and per‑service DynamoDB/Redis stores, reaching about 200 functions and boosting deployment cadence from 4–6 to 80–100 deployments/month. Replaces Mixpanel ($3,000/month) with BigQuery/Athena streaming for ad‑hoc analytics costing only cents, and implements fan‑out for timelines via Kinesis→Lambda→SNS→Redis with Step Functions for intra‑domain workflows and synchronous fallbacks where immediate consistency is required.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 1:47Monolith to serverless
- 5:44Event-driven architecture
- 7:30Strangler pattern
- 12:24Search feature rebuild
- 14:08Mixpanel to BigQuery
- 17:39Real-time timeline
- 21:51User recommendation system
- 24:20Serverless benefits
- 25:46Q&A
related talks
44mHow to build a social network entirely on serverless
33mHow to build a social network entirely on serverless
39mServerless in production, an experience report
01h 02mServerless in production, an experience report
01h 02mFrom Monolith to Serverless
47mServerless in production, an experience report
41mTechnology mix: Around the serverless in 180 days
30mEvolving Serverless Architectures
