Monitoring OpenShift - Upgrade from version 3 to 4
- Instead of deploying one DaemonSet our configurations deploy 2 DaemonSet and 1 Deployment. One DaemonSet with configurations specific for Masters, second DaemonSet with configurations specific for Nodes (non-Masters). And one Deployment, which collects data from the whole cluster.
- DaemonSet is deployed within host network, to be able to connect to system components and collect metrics from them.
- All workloads and configurations are moved under collectorforopenshift Project.
- Process and container metrics have a different format, which makes it impossible to use collector v4 with Splunk application v3. After migrating all components to v4, you should be able to see all data, including data, collected before the upgrade.
- ClusterRole includes more endpoints, to be able to collect metrics exported with Prometheus format.
- ConfigMap consists of a group of configurations, which allows mapping parts of configuration to used workloads.
- Before we showed default values of the configuration under commentaries. Now all the default values need to be explicitly defined in the configurations.
Upgrade Splunk application
Download version 4.0 from SplunkBase.
Delete currently deployed version of collector for openshift.
oc delete daemonset collectorforopenshift --namespace default oc delete configmap collectorforopenshift --namespace default oc delete clusterroleBinding collectorforopenshift --namespace default oc delete clusterrole collectorforopenshift oc delete serviceaccount collectorforopenshift --namespace default
Download latest Configuration Reference to be able to
upgrade to version 4 of
Follow installation instructions Monitoring OpenShift Installation