Kubernetes pattern (Discovery and Service Mapping - 1.0.90) merges Kubernetes Cluster CIs because it fails to populate IP addresses - Known Error

Kubernetes pattern (Discovery and Service Mapping - 1.0.90) merges Kubernetes Cluster CIs because it fails to populate IP addresses Description Kubernetes pattern (Kubernetes Leader shared library) merges Kubernetes Clusters CIs when https:///namespaces/kube-system/endpoints/kube-controller-manager is not exposed.

When it happens (it looks like this is the case for all recent Kubernetes releases by default) the failover procedure fails and it doesn't populate cmdb_ci_kubernetes_cluster.ip_address due to bug in the IP address / Hostname resolution section.

Resolving IPs does't make sense because DNS reverse entries are not which means IP resolution will fail.

Steps to Reproduce

Provide Kubernetes URL which contains IP address in the URL and run discovery If Kuberentes pattern fails to populate IP addresses multiple Kubernetes Cluster CIs will be merge into one CI.

Workaround Kubernetes shared library has been updated. The fix has been targeted for September 2022 which means you will have to check if your Kubernetes Leader shared library already contains it.

Open Kubernetes pattern in the Pattern Designer and expand Kubernetes Leader and check if it contains any step from the list below.

If it does and you still have an issue please submit a case and Support will help you.

Check if IP Address step has been introduced because customers could provide Kubernetes URLs which contain either name or IP address Copy existing IP from host variable step copies IP because it doesn't make sense to resolve IP again. Resolve primary_node_name step uses DNSUtils now instead of the error prone nslookup Check if IP address is empty and log warning step checks if host_ip has been populated. If that's not the case it will add a log entry and continue. ( DNS does not resolve Kubernetes cluster hostname! ServiceNow will merge Kubernetes Cluster CIs if their IP addresses and ports match ) There is always an option to apply the attached update ( CSTASK321776 - update_set - kubernetes leader - cluster master node ip.zip ) by following the procedure.

Update Discovery and Service Mapping Patterns if you haven't already. If your version is greater than 1.0.90 please check if it already contains the fix. Extract "CSTASK321776 - update_set - kubernetes leader - cluster master node ip.zip" Import sys_remote_update_set_abe23cb01b8f8510a95564a1b24bcb63.xml only (Discovery and Service Mapping Patterns application) Sync MID servers Re-run discovery Related Problem: PRB1583186