April 2015 Summaries

The blog post, authored by Gianluca Borello, delves into the intricacies of troubleshooting networking issues and configuration management in CoreOS environments using Sysdig. It highlights the challenges associated with flannel networking, which enables direct communication between containers across different hosts by creating a virtual network through UDP tunneling. The article provides an example of how incorrect network configuration can lead to routing issues, using Sysdig to diagnose and resolve these problems. It also covers troubleshooting confd, a daemon that syncs configuration files with values from etcd, demonstrating how Sysdig can monitor real-time file changes and detect configuration errors due to race conditions in fleet unit files. The post underscores the complexity of managing CoreOS and containerized systems and promotes Sysdig as an essential tool for gaining visibility and resolving issues in such environments.

The text discusses the integration of Sysdig with CoreOS environments, highlighting the unique visibility Sysdig provides for troubleshooting and observing transactions within containerized applications. The article explains the deployment of Sysdig on CoreOS using a Docker container, allowing for full visibility of system events across containers without the need for additional instrumentation. It provides a step-by-step guide on setting up a basic CoreOS infrastructure using Fleet unit files to manage frontend and haproxy containers, which are configured with the help of etcd and confd for service discovery and load balancing. Various use cases are demonstrated, such as observing application transactions and commands executed within containers, showcasing how Sysdig can aid in troubleshooting containerized applications. The text concludes by mentioning the Sysdig Cloud solution for unified monitoring and alerting in distributed CoreOS environments, emphasizing the ease of use and passive nature of Sysdig in providing deep insights into container activities.