Cancel, Retry and Timeouts: Keep Your Sanity Thanks to Reactive Programming
summary
Reactive pipelines treat cancel, timeout and retry as first-class composable concerns: subscribing a Mono/Flux returns a disposable from subscribe that cancels the entire composed flow, a single timeout operator can limit the whole pipeline or per-inner call depending on placement, and retries use exponential backoff with jitter to avoid retry storms. Demo code shows parallel inner requests that get cancelled or time out, and configurable retry with backoff and jitter that limits attempts, preserves cancelability, and composes cleanly with timeouts.
AI-written summary · about
chapters (9)
- 0:00Introduction to talk
- 0:44Imperative blocking requests
- 3:54Cancellation with futures
- 12:31Reactive cancellation with Reactor
- 21:10Timeout challenges in APIs
- 26:08Reactive timeouts with Reactor
- 37:31Retry strategies explained
- 40:42Reactive retries with Reactor
- 48:06Summary and Q&A
related talks
42mConquering Time with Functional Reactive Programming
01h 28mIntroduction to Reactive Programming
41mBuilding Reactive Pipelines with Kotlin & Spring
53mReactive Mythbusters With Reactor
48mReactive Streams - Beware of Dangers!
48mReactive for the Impatient
38mCuring your event processing blues with Rx
47mReactive Spring