How one team built their first microservice
summary
A small microservice that handles goal comments and a live feed is built using dotnet core, Docker containers deployed to Docker + AWS ECS with Terraform, DynamoDB for state, SQS-style queues for async messaging, CQRS and read-store projections, plus an anti‑corruption layer to keep the monolith isolated. Cycle time rises from ~4 to ~8 days during the transition then falls to ~2 days once stable. Persistent message loss (network, bugs, retries) causes lost updates — the team concludes event sourcing would have made repairs far simpler.
AI-written summary · about
chapters (11)
- 0:00Start
- 7:06Value Stream Map
- 12:32Decoupling
- 18:17Downsides
- 19:28Move to Continuous Deployment
- 21:40Jquery
- 25:26How a Continuous Deployment Pipeline Works
- 28:29View Model
- 30:22Anti-Corruption Layer
- 50:54Docker Compose
- 52:40How Did You Choose Which Version of Net Core To Use and and Were You Using the Latest All the Time
related talks
10mMicroservices, Messaging, and Failure: How Simplifying Our Architecture Saved Us
23mOn the road from Monolith to Micro
01h 01mMicroservices Gone Wrong
56mMicroservices from the trenches
22mMicroservices Highway to hell
24mMicroservices in Practice
38mPatterns and Practices for Real-world event-driven microservices
41m.NET Microservices in Real Life