From Legacy to High-Load: Evolution of a Web Application
summary
A legacy PHP/Apache/MySQL monolith was migrated into a microservice architecture using nginx + php‑fpm, ElasticSearch for fast search, Redis for sessions, RabbitMQ for queues, a Galera DB cluster, and containers. Concrete incidents and results are described: a 4‑hour outage during ticket sales was resolved by adding a missing database index; Eurovision finals handled ~3,200 concurrent page users with a 12k–52k queue; Imagine Dragons created ~52,500 queued requests and ~$1M in six hours. The team also developed a storage service, a Crossbar RPC core, and host‑process workers that proved to scale.
AI-written summary · about
chapters (6)
- 0:00Start
- 7:25INITIAL PROJECT ARCHITECTURE
- 9:08LOAD IN 2012
- 27:59TRANSFORMATION: PHASE 1
- 32:49TRANSFORMATION: PHASE 2
- 42:14IMAGINE DRAGONS 2018
related talks
56mLarge scale websites performance optimisation tricks
44mIntroduction to application scaling
46mHow to escape Legacy Hell
50mBuilding and Operating Online Shops under Extreme Load
30mBringing it all together
36mApache, nginx, Docker and the Swear Words We Love
18mHigh-Load Systems: Challenges in Social Network Development
45mScaling: from 0 to 20 million users