Skip to content

Latest commit

 

History

History
40 lines (30 loc) · 1.59 KB

README.md

File metadata and controls

40 lines (30 loc) · 1.59 KB

Cluster Monitoring

Install

Prometheus and Graphana tools are used to monitor the cluster. Install as given below:

  1. Select 'Monitoring' App from Rancher console -> Apps & Marketplaces.

  2. Select Chart Version 100.1.3+up19.0.3 from dropdown in Rancher console -> Apps & Marketplaces.

  3. In Helm options, open the YAML file and disable Nginx Ingress.

  4. Click on 'Install'.

Prometheus

All MOSIP modules have been configured to let Prometheus scrape metrics.

Grafana

To load a new dashboards to Grafana, sign in with user and password from rancher-monitoring-grafana in cattle-monitoring-system namespace of Rancher cluster.

Important default dashboards:

Grafana dashboard Description
Rancher/Cluster (Nodes) Consolidated view of all nodes
Rancher/Node View of each node
Kubernetes/Pesistent Volumes Storage consumption per PV
Kubernetes/Compute Resources/Workload Resources per deployment/statefulset
Kubernetes/Compute Resources/Cluster Resources namespace wise

To see JVM stats you may import chart number 14430 in Grafana dashboard.

JVM stats

MOSIP pods make JVM stats availalbe for Prometheus to scrape the pods. The typical endpoint looks like <base url>/actuator/prometheus. The scrapping is enabled via module Helm chart (see metrics section of values.yaml).

A sample of metrics that is pulled by Prometheus is given in pod-jvm-scraped-metrics-sample.txt

Inodes

To see free inodes on a particular node, login to the node an run

df -ih