Explore the intricacies of building a distributed task scheduler using Akka, Kafka, and Cassandra in this 37-minute conference talk from Strange Loop. Dive into PagerDuty's journey of upgrading their microservices infrastructure to handle dynamically scheduled tasks for incident alerts, on-call notifications, and administrative chores. Learn how the Core team leveraged the strengths of these three technologies to overcome challenges in resilient, distributed task scheduling. Discover real-world solutions to critical issues such as dynamically adjusting for increased task load with zero downtime, guaranteeing task ordering across multiple servers, ensuring task execution during datacenter outages, and preventing task duplication. Gain insights into distributed system concepts applied in a practical context, including dynamic load balancing, datacenter failover strategies, and maintaining task order integrity. While some familiarity with distributed queueing and actor systems is beneficial, this talk offers valuable lessons for anyone interested in robust, scalable microservices architecture. Read more