Black Friday Brilliance: Managing a Billion Transactions with Tech, and Teamwork
summary
Claims concrete: the platform processed over 1 billion API requests in 2023 with 99.99% availability, traffic spikes ~70% above baseline and 100x growth over a decade. Infrastructure uses Google Cloud (Pub/Sub, BigQuery, Cloud monitoring), a mix of VMs and Kubernetes, a Pub/Sub queue and caching layer, and a move from Windows VMs to Linux to cut costs from ~$300k to ~$250k/month. Capacity testing starts three months out, data-center failovers are rehearsed, and microservices start in under a minute after reducing startup from ~10 minutes.
AI-written summary · about
chapters (14)
- 0:00Introduction and agenda
- 1:16Black Friday history and scale
- 3:43Era of cloud migration
- 5:5112 factor app methodology
- 9:01Who is Locate and Black Friday impact
- 9:43Preparations three months ahead
- 11:10Final preparations before Black Friday
- 12:36Current tech stack overview
- 14:21Traffic balancing and legacy systems
- 15:25Past outages and solutions
- 16:49Future tech stack improvements
- 18:13Multi-cloud strategy and data centers
- 19:35Reducing cloud costs
- 20:19Key takeaways and conclusion
related talks
46mBlack Friday Brilliance: Managing a Billion Transactions with Tech, Tactics
42mBlack Friday Brilliance: Managing a Billion Transactions with Tech, Tactics, and Teamwork
33mHigh availability by offloading work
56mGetting all the 99.99(9) you always wanted
10mBuilding High-Performance Financial Data Pipelines
27mScalable by Design: How to Build an Elastic Cloud Infrastructure...
24mBuilding Planet-Scale Payment Platforms
28mNavigating the Transition: SRE Challenges and Highlights