Monitoring Kubernetes Cluster MenggunakanPrometheus dan Grafana

Abstract

The large number of client requests can cause excessive workload on the server so that it can

make the server down. Because it uses cloud computing services that can increase the number of client

requests. With the cluster technology in cloud computing, the server workload will be divided evenly or

balanced to each server. Clustering can be combined with containers where each process or application

running on each container has the same kernel. Therefore, we need an application capable of managing

containers, one of which is Kubernetes. Kubernetes has several components, namely clusters, pods, services

and nodes that need to be monitored. To be able to carry out monitoring, an application that can help is

used, namely Prometheus and Grafana. Prometheus will retrieve the data in Kubernetes, then the data that

has been obtained by Prometheus can be visualized with Grafana. Grafana can convert metric data into

graphs that are easy to understand and interactive. Based on the results of functionality testing that has

been carried out, Prometheus has succeeded in reading data from the target server and Grafana has

succeeded in displaying it in graphical form. In the test, from 5 trials of 20-100 user access, the monitoring

system can show the amount of CPU usage, memory, and traffic load that can continue to increase

according to the user's access load.