"ASYNC: Discovery - MultiPage Sensors Pattern Launcher: Kubernetes - MultiPage 74" is consuming memory around 1.1GB - Known Error

"ASYNC: Discovery - MultiPage Sensors Pattern Launcher: Kubernetes - MultiPage 74" is consuming memory around 1.1GB Description "ASYNC: Discovery - MultiPage Sensors Pattern Launcher: Kubernetes - MultiPage 74" is consuming memory around 1.1GB. This happens on the last page of the result, so it is a deletion strategy, so it should be improved to use script or try to make the pattern LP. Kubernetes schedule gets too many results and accumulates to causing the deletion strategy to use too much memory. https:// .servicenow.com/now/nav/ui/classic/params/target/sa_ci_to_pattern_list.do%3Fsysparm_query%3Dpattern%253D96324ec09f3303002ae692ec757fcf61%26sysparm_first_row%3D1%26sysparm_view%3D

Steps to Reproduce

"ASYNC: Discovery - MultiPage Sensors Pattern Launcher: Kubernetes - MultiPage lastpage" is consuming memory around 1.1GB

Workaround The issue is fixed in Discovery and Service Mapping March 2026.

The workaround is given in the update sets and it should significantly reduce performance load.

PRB1956848_DeletionStrategyChange.xml

PRB1956848_K8SPattern_Changes.xml

PRB1956848_Kubernetes_OOM_issue.xml

What Gets Deleted Kubernetes records are deleted or marked as absent based on deletion strategy configured in sa_ci_to_pattern table.

Do any actions needed after upgrade to the fixed version? No action need from customer side on upgrade to the fixed version

Changes In Update set:

Changes in PRB1956848_DeletionStrategyChange.xml Update set:

For the following Kubernetes workload CIs deletion strategy is set to Mark as Absent from Delete since deletion operation takes more time and on large cluster is not performance efficient. cmdb_ci_kubernetes_deployment cmdb_ci_kubernetes_daemonset mdb_ci_kubernetes_replicaset cmdb_ci_kubernetes_replicationcontroller Deletion strategy script is enhanced and made sure its memory efficient

Changes in PRB1956848_K8SPattern_Changes.xml Update set:

we should not create relationship between image and workload CIs since this relationship exists via pod and image.

In this update set we disabled following steps:

Create relations of deployments to image

Create relations of statefulset to image

Create relations of daemonset to image

Create relations of replicaset to image

Create relations of replicationcontroller to image

Changes in PRB1956848_Kubernetes_OOM_issue.xml Update set:

Memory Optimization changes are added to the pre script sa_pattern_prepost_script_c63842411be1711066d1c880604bcbc9.xml

Keep original payload string to avoid re-stringify overhead var originalPayload = payload;

The variable payload = null; // Release reference to help GC

Release payloadObj reference before return; payloadObj = null;