Environment
CentOS Linux release 7.8.2003 (Core)
:
Following servers are running with 192.168.1.151 eternal1 eternal1.OracleByExample.com 192.168.1.152 eternal2 eternal2.OracleByExample.com 192.168.1.153 eternal3 eternal3.OracleByExample.com 192.168.1.191 PromGraf PromGraf.OracleByExample.com
Apache Cassandra
3 node Cassandra cluster cluster_name: 'id_cluster'
version : Apache Cassandra 4.0-beta4
running on following servers :
192.168.1.151 eternal1 eternal1.OracleByExample.com 192.168.1.152 eternal2 eternal2.OracleByExample.com 192.168.1.153 eternal3 eternal3.OracleByExample.com
We will configure Cassandra's built in metrics-reporter
to extract and publish metrics to Prometheus.
We will also configure Prometheus node_exporter
on each of our Cassandra nodes to extract and publish metrics to Prometheus.
Prometheus & Grafana
We will configure and run Prometheus & Grafana
on following server :
192.168.1.191 PromGraf PromGraf.OracleByExample.com
Prometheus
will gather and organize all collected metrics into its internal time-series database. Grafana
will consume the metrics from Prometheus and display them in a nice dashboard.
Prometheus Alertmanager
In future, we should implement Alertmanager