Microservices enable scalable cloud-native systems but also introduce continuous failure modes such as timeouts, retry storms, and cascading outages. This session distills evidence-based recovery patterns from a PRISMA-aligned review of 26 peer-reviewed studies (2014–2025), covering techniques including circuit breakers, retries with jitter and budgets, bulkheads, sagas with compensation, idempotency, backpressure, and chaos engineering. It highlights real-world trade-offs across tail latency (P95/P99), availability, resource overhead, and operational complexity, and provides actionable guidance for designing resilient, self-healing microservice architectures. Based on an IEEE-accepted conference paper presented at ICICyTA 2025.

https://youtu.be/S1Ed_EV4zSc