Microservice architectures involve a large number of highly decoupled lightweight apps - leading to a dense web of network calls.
As such overall response time even for a single complete flow depends directly on the time spent in the involved service-service interactions.
Via selective replication & options like k8s federation, users can deploy app replicas to reduce the latency impact.But:

How to monitor these latencies?
What all (apps) & where all (nodes) to replicate?
What about complex and dynamically changing interactions patterns?

Scalable? Certainly Not!

In this talk, Yash will present a smart & lightweight approach to automatically grasp the app flows, interactions & latency, and tell precisely which all nodes to replicate the apps. Conversely, if possible, instead of replication, how to perform clever migrations to save infra costs & still reduce overall system latency. Yash will highlight the results from a real world cluster & how you can use a similar approach for yours.