Cloud-native environments often promise high availability, but real production systems still fail in ways that managed services and standard patterns alone can’t prevent.
This session digs into the real failure modes encountered in large-scale distributed systems, from cascading retries and dependency latency spikes to partial outages and resilience gaps that only surface under real load.

Through practical case studies and insights, we’ll explore resilience patterns that truly work in practice, including circuit breakers, bulkheads, graceful degradation, and idempotent workflows, along with common anti-patterns we learned the hard way.
Attendees will follow real scenarios to observe how these patterns behave when things go wrong, how observability helps uncover hidden failure chains, and how small configuration choices can significantly affect system resilience.