Operations and Lifecycle
Domain 6 covers health, backup, upgrade, and troubleshooting across the OpenShift and Cloud Pak for Data layers.
Monitoring
Start at the platform layer: check the control plane and operators, then service instances, then pods, events, and persistent volume claims. cpd-cli manage status summarizes the instance; OpenShift monitoring and logging provide metrics and logs. Alert on certificate expiry, PVC usage, and failed jobs, not just pod restarts.
Backup and upgrade
Back up persistent volumes and platform configuration before changes. Follow the supported upgrade path and check the version compatibility matrix for OpenShift, Cloud Pak for Data, and each service. Upgrades are staged: operators first, then the control plane, then services.
Worked example
After an upgrade, a service pod is in CrashLoopBackOff. Do not reinstall immediately. Check events for image pull or permission errors, confirm the PVC is bound, inspect the operator logs, and verify the service version is supported on the new control plane. The reasoning separates infrastructure symptoms from application configuration.
Traps
Troubleshooting only the pod while ignoring the operator. Upgrading a service before the control plane. Skipping backups because storage is replicated. Assuming a running pod means a healthy service; synthetic checks matter.