Open Source Summit + Embedded Linux Conference Europe 2020

Since the advent of containerized infrastructure, one can’t simply extend the monitoring strategy that worked for VMs and expect it to work for containers. Due to container portability and the rise of Kubernetes, the need to adopt multi-cloud strategies has added even more complexity to application monitoring. The main reason is that Kubernetes adds another multi-component layer to software environments that must be monitored. One has to keep track of the health of the cluster, pods, containers, storage, and networking components within the cluster along with monitoring the applications and the underlying infrastructure itself. The application as a whole, Kubernetes itself, and the infrastructure can have issues under the unforeseen circumstances. To stay ahead of these issues, there is a need to have a comprehensive monitoring regime that addresses all the layers of a containerized, Kubernetes based environment. This proposal talks about taking a three-tiered approach to Kubernetes monitoring by identifying key pillars to adopt in order to monitor workloads most effectively with the use of open-source tools like Grafana, Prometheus, ELK and Cortex.