Dittyboppter is a quick-and-dirty way to deploy system-level submetric monitoring with Grafana in an OpenShift 4 environment. It deploys a stand-alone transient Grafana pod with default dashboards to monitor things like CPU, memory, network, and disk activity. There are also dashboards for monitoring rook-ceph components, and the deployment can be customized with your own dashboards. The Grafana charts expect to receive metrics from an existing Prometheus deployment and node exporters.

Right now Dittybopper has a number of FIXMEs that need to be addressed before it will be more portable across k8s/OpenShift environments. It should generally deploy out-of-the-box with OpenShift 4. Other environments will likely have a prerequisite to first stand up a Prometheus pod, and the Dittybopper scripts and templates will need adjustment accordingly. The default dashboard included at dashboards/defaults/dittybopper-master_nodes.json is currently designed for a 3-node OpenShift 4 deployment with converged rook-ceph storage on the master nodes.

Pull requests are encouraged. If you find this tool useful, please help extend it for more use cases.

$ git clone https://github.com/cloud-bulldozer/dittybopper.git
$ cd dittybopper
$ ./deploy.sh [-c <kubectl_cmd>] [-m <host1> <host2> <host3>] [-n <namespace>] [-p <grafana_pwd>]

See ./deploy.sh -h for help.

Simply running ./deploy.sh with no flags will assume OpenShift, the dittybopper namespace, and admin for the password.

This will import a dashboard (json) into an existing Dittybopper Grafana deployment.

$ ./deploy.sh -i <path_to_dashboard_json_file>

$ ./deploy.sh -d