KRKashif Rezainarchitecture-in-practice.hashnode.dev·19h ago · 7 min readHandling Duplicate Events Safely with IdempotencyIn the previous article, we looked at retries and DLQ. Retries help you recover from temporary failures. But they also create a new risk: The same event may be delivered, retried, or processed more t00
KRKashif Rezainarchitecture-in-practice.hashnode.dev·1d ago · 8 min readRetry Strategy in Event-Driven Systems: Resilience or Retry Storm? Retries usually start with good intent. Typical transient failures in distributed systems A network call times out.A database becomes slow.A downstream API is unavailable. These are not edge cases. T00
KRKashif Rezainarchitecture-in-practice.hashnode.dev·5d ago · 10 min readClosing the Gap Between DB Commit and Event Publishing with the Outbox PatternHow the Outbox Pattern makes event publishing durable after a business transaction succeeds In the previous article, we saw a dangerous failure window: Save business data Commit DB transaction Publish00
NSNilay Shahaneincuriousnilay.hashnode.dev·Aug 15 · 8 min readI Tried Designing a Distributed Rate Limiter. Here’s Every Place It Broke.Somewhere out there is a client with a bug — a while (true) loop with a missing break condition, hammering /api/order/create as fast as the network allows. while (true) { await fetch('/api/order/cre21T
