Herbert Abdillah

Jan 28, 2021

1 min read

Easy Monitoring + Logging Stack in Kubernetes Cluster With Helm

There are many helm chart available on internet. I prefer to use from https://artifacthub.io/ because i see the stack always updated/maintained. From there i found 2 most used choice : ELK Stack, and Loki + Prometheus + Grafana Stack.

In ELK Stack, filebeats used to collect log from pod that stored in node filesystem. We may use logstash if we want transform the log. The log will be send to elasticsearch. Kibana will use visualization data that originated from elasticsearch. For installation, we must install stack and configure filebeats, logstash (optional), elasticsearch, and kibana one by one. Although the configuration is easy enough. Just edit the values.yaml according to our needs.

In Loki + Prometheus + Grafana Stack. Same like above, Promtail used for log collecting, data saved in loki, and visualized in grafana. For the monitoring, it used kube-state-metrics and prometheus-node-exporter. For Installation, we just install and do a little config twice. First, install the kube-prometheus-stack. Second, install the loki-stack without grafana dashboard (edit in values.yaml). Third, we use grafana dashboard in kube-prometheus-stack, and add loki datasource for logging.

So i think the Loki + Prometheus + Grafana stack is easier.