As described in the the Accessing Kubernetes logs locally section, basic logging can be used to detect configuration problems, but for cluster-level logging, an external backend is required to store and query logs. A cluster-level logging stack can help you quickly sort through and analyze the high volume of production log data that's produced by your application in the Kubernetes cluster. One of the most popular centralized logging solutions in the Kubernetes ecosystem is the Elasticsearch, Logstash, and Kibana (ELK) stack.
In the ELK stack, Logstash is used as the log collector. Logstash uses slightly more memory than Fluent Bit, which is a low-footprint version of Fluentd. Therefore, in this recipe, we will use the Elasticsearch, Fluent-bit, and Kibana (EFK) stack. If you have an application that has Logstash dependencies, you can always replace Fluentd/Fluent Bit with Logstash.
In this section, we will learn how...