December 2015 Summaries

The blog post explores the challenges and solutions associated with troubleshooting Kubernetes services using sysdig, focusing on scenarios where application failures occur due to misconfigurations. It begins with a detailed examination of how a Kubernetes service operates under normal conditions, including DNS resolution by SkyDNS and request forwarding via the Kubernetes proxy. The article then delves into specific troubleshooting scenarios, such as incorrect port configurations, label mismatches, firewall misconfigurations, and DNS issues when accessing services from the host. Through examining system events and proxy behavior, the author provides insights into identifying and resolving these issues, highlighting the importance of understanding expected system behavior for effective troubleshooting.

In this first part of a two-part series, Gianluca Borello delves into the workings of Kubernetes services using the open-source tool Sysdig. Borello explains that a Kubernetes service acts as an abstraction layer that provides a single entry point for accessing a group of pods, functioning like a dynamic load balancer. This simplifies application development by eliminating the need for manual service discovery mechanisms. He sets up a simple Kubernetes environment with two nginx pods and a service, demonstrating how these services use DNS extensions to map service names to virtual IPs, allowing transparent access by other pods. The article explores the technical intricacies of how Kubernetes handles DNS requests and routes them through a virtual IP using Linux kernel Netfilter capabilities, ensuring stable service discovery even if pods change locations. Borello highlights how Sysdig can capture and analyze network events to illustrate these processes, setting the stage for troubleshooting in more complex or faulty environments in the upcoming second part.