Analysing and Alerting on Application Logs Within Kubernetes Infrastructure

EasyChair Preprint no. 10457

Abstract

Software applications designed as a set of micro-services constantly require supervision. Logging of applications can help development teams explore and discover new insights in the workflow within micro-services, in addition to errors, exceptions, mistakes, and inefficiencies. Kubernetes is the most popular technology to maintain, manage and work with micro-services. Nevertheless, Kubernetes technology lacks in ability to effectively and efficiently provide logging information to the end-users. It is worth mentioning that Kubernetes has its logging tools, however, they are inefficient and not representative to obtain useful and insightful information. To cope with this, there exist various solutions to provide powerful logging capabilities. However, in this paper main focus will be on two of the most popular stacks of logging solutions that can also be integrated into Kubernetes infrastructure, namely Elastic stack and PLG stack with Prometheus. In addition to logging analysis, it is also necessary to have an opportunity to alert on these logs based on specified rules, e.g. alert on errors. The goal of this research paper is to understand which stack is the most suitable for providing application logging within Kubernetes infrastructure based on chosen metrics for a Java logger application, as well as define which stack is more relevant in terms of implementation based on a literature review.

Keyphrases: Docker, Elastic Search, Elastic Stack, ELK Stack, Grafana, K8s, Kibana, Kubernetes, logging, Logstash, Loki, PLG stack, Prometheus, Promtail

@Booklet{EasyChair:10457,
  author = {Danila Koryugin},
  title = {Analysing and Alerting on Application Logs Within Kubernetes Infrastructure},
  howpublished = {EasyChair Preprint no. 10457},

  year = {EasyChair, 2023}}