October 2022 Summaries

Kubernetes logging can be challenging, especially at scale. Traditional solutions like ELK (Elasticsearch Logstash Kibana) have limitations when dealing with large volumes of log data from distributed environments. Loki, a cloud-native log aggregation system, offers a more efficient and cost-effective solution for managing logs in Kubernetes environments. It provides features such as labels, cloud-native backend storage, and a query language called LogQL, which are inspired by Prometheus' PromQL language. Loki's architecture is designed to distribute the load, protect the system from attack, and make use of efficient storage mechanisms. The "PLG" stack, consisting of Loki, Promtail, Grafana, and Tempo, provides a comprehensive observability solution for Kubernetes environments.

The Kubernetes ecosystem includes a few complementary add-ons for aggregating and exposing monitoring data from your Kubernetes cluster, but one of these useful add-ons is the Metrics Server. However, its installation can introduce issues such as maintenance requirements and potential crashes. The author's company, groundcover, aimed to expose K8s node resources and usage metrics for their customers without installing a 3rd party tool like the Metrics Server. They found that the Kubelet, which runs on each node of a Kubernetes cluster, already exposes resource metrics through APIs. By querying these APIs, they can get the job done without deploying a Metrics Server, making it a lightweight and frictionless approach to Kubernetes application monitoring.