Kotlin Backend Development 6 Yrs Recap. The Good, the Bad and the Ugly
summary
Kotlin runs production microservices at a $4B startup: 1.2M LOC, 150 backend developers and 30+ microservices on AWS/Kubernetes using Dropwizard, Gradle and Jackson Kotlin module. Key wins: 6,000 data classes, concise immutability and workable coroutine MDC propagation for tracing. Pains: 11,000 invocations of the null-bypass (!!) pattern, risky extension functions and type aliases, Kotlin-script compatibility for deployed rating formulas, and ongoing headaches upgrading Gradle/bytecode and Java/Kotlin versions.
AI-written summary · about
chapters (19)
- 0:00Start
- 8:45Kotlin Version
- 13:47Data classes - The Prince that was Promised
- 15:56Data Classes - Jackson Serialization
- 17:21The Promise of Null Safety
- 27:07Extension Functions. my preferance
- 28:32Our concurrency model & Tx Ctx
- 30:44Coroutines: launch
- 33:00Type Aliases. The feature we use too much
- 33:30Type Aliases. Are not Type Safe
- 33:57Type Aliases. What is behind the magic
- 35:48Put Type Aliases to Good Use
- 36:35Inline Classes to the rescue
- 38:16Migrations along the Years
- 40:51Kotlin Java Bytecode Compatibility
- 41:35Gradle - Java bytecode Compatibility
- 41:47Kotlin API deprecation
- 43:29Detaching gradle compiler version
- 44:16Build Time
related talks
40mDeveloping Micro Services With Kotlin
46mKotlin Is for Hipsters
44mKotlin is for hipsters
01h 05mKotlin: why and how?
52mKotlin - Ready for Production
59mKotlin - Ready for Production
38mKotlin from Zero to Production: The C6 Bank Experience
48mKotlin for the back-end – Does the fancy code bring much value?