Scaling Production data across Microservices

on YouTube ↗27m

" summary

Describes Fiverr’s migration from a LAMP monolith to ~200 microservices and the data-sharing problems that arose when data moved behind REST APIs (performance, reliability, team dependencies). Explains a concrete solution using Kafka as a central data hub, combined with patterns from CQRS and Domain‑Driven Design to share production data across services and enable faster iteration.

AI-written summary · about

" chapters (12)

1. 0:00Start
2. 1:44Trouble maintaining the same cadence
3. 2:25monitor for errors and performance issues
4. 2:54harder to upgrade our 3rd party dependencies
5. 3:39Model View Controller
6. 4:01Use the latest tech outside the Monolith
7. 4:46complex deploy sequences involving the Monolith and several Microservices
8. 11:13Can we avoid the coupling to the source database?
9. 21:08orchestration vs choreography • shift to choreography
10. 23:45compete order by user business flow
11. 24:56key take-aways: . Practice evolutionary architecture • Respect the Data Dichotomy
12. 25:54key take-aways: . State Change Events Read optimized views decouple querying patterns and increase performance . They are not mutually exclusive and their Ordering is important!

" related talks

• 44mA Software Engineer's Journey Through Microservices Data SharingDaniil Doronkin · DevDays Europe 2024
• 53mHow do we build high‑load microservicesAnton Moldovan · .NET Fest 2017
• 36mMicroservices Lessons LearnedSusanne Kaiser · Voxxed Days Microservices 2018
• 54mFrom a distributed monolith to a microservices solutionJan de Vries · NDC Sydney 2020
• 49mChallenges in Implementing MicroservicesFred George · GOTO Amsterdam 2015
• 29mMoving from A and B to ~150 microservices, the journey and learningsGeshan Manandhar · Container Camp AU 2019
• 50mChallenges in Implementing MicroservicesFred George · GOTO Copenhagen 2015
• 25mData-Driven Migration from Monoliths to Event-Driven MicroservicesAmlan Ghosh · Site Reliability Engineering (SRE) 2025