In the world of distributed computing, containers, and microservices, a lot of the interactions and communication between services is done via RESTful APIs. While developing these APIs and interactions between services, I often have the need to debug the communication between services, especially when things don’t seem to work as expected.
Before the world of containers, I would simply deploy my services on my local machine, start up Wireshark, execute my tests, and analyze the HTTP communication between my services. This for me has always been an easy and effective way to quickly analyze communication problems in my software. However, this method of debugging does not work well in a containerized world.
First of all, the containers most likely run on an internal container platform network that is not directly accessible by your machine. A second problem is that, in compliance with container design best practices, containers contain only the minimal set of applications and libraries needed to execute their task. This means that a tool like
tcpdump is usually not available in a container. This makes debugging and analyzing network traffic between containers and, thus, debugging of inter-microservice communication a bit harder than in the non-containerized world. This article shows one solution.
Continue reading “Using sidecars to analyze and debug network traffic in OpenShift and Kubernetes pods”