v_cmdb_running_process ml_solution stuck with Update State = "Error while updating Solution"Issue The ML Readiness page displays a warning indicating that the Application Fingerprinting or Connection Suggestion training job is stuck. The v_cmdb_running_process ml_solution record shows an Update State of “Error while updating solution”. Internal investigation confirms no issues on the ML training server side. Because the solution is of type application_suggestion, the “Update and Retrain” option is restricted in the UI and, requiring resolution via background script.Symptoms• ML Readiness page displays a warning for Application Fingerprinting or Connection Suggestion. • The ml_solution record for v_cmdb_running_process shows Update State = “Error while updating solution”. • Records in the ml_cluster_detail table are not being created or updated. • The solution remains stuck and does not self-recover.Facts• Affected ML solution: v_cmdb_running_process (application_suggestion type). • The “Update and Retrain” button is visible in the UI; however, clicking it for an application_suggestion solution will throw the error: “Retraining is not allowed for application suggestion clustering solutions.” • The solution update state must be reset manually via a background script. • The system property glide.app.fingerprint.maxprocess.update controls the update set size for processing. • Reducing the update set size to 10,000 records allows the update job to complete in smaller, faster batches. • The scheduled job “Applications suggestion - ITOM Autodiscovery” is responsible for triggering ml_solution updates for the application_suggestion solution type. • Once the Update State is reset to solution_complete, the scheduled job will detect the change and initiate a new update cycle, resetting the state to 0%. ReleaseNot a version specific issue.CauseThe ml_solution record for the application_suggestion solution type (v_cmdb_running_process) enters an error state during the update process. The update set size being processed is too large, causing the update job to fail and the solution to become stuck in the “Error while updating solution” state. Because the solution type is application_suggestion, the standard “Update and Retrain” UI option is restricted, preventing a straightforward recovery through the interface.ResolutionPerform the following steps in order to recover the stuck ml_solution and prevent recurrence. Step 1 — Reset the ml_solution update state via background script Navigate to System Definition > Scripts - Background and execute the following script, replacing the placeholder with the sys_id of the application_suggestion ml_solution record: var sol = new GlideRecord("ml_solution"); sol.get("<solution_sys_id>"); // sys_id of the application_suggestion solution var state = "update_state"; sol.setValue(state, "solution_complete"); sol.update(); Step 2 — Create the system property to reduce update set size Navigate to sys_properties and create a new system property with the following value: glide.app.fingerprint.maxprocess.update = 10000 Setting this value to 10,000 reduces the batch size processed per update cycle, allowing each cycle to complete faster and preventing the solution from timing out or erroring. Step 3 — Execute the Applications suggestion - ITOM Autodiscovery scheduled job Navigate to System Scheduler > Scheduled Jobs and locate the job named "Applications suggestion - ITOM Autodiscovery" Execute this job manually. This scheduled job is responsible for triggering updates to the application_suggestion ml_solution. When it detects that the update state is solution_complete, it will initiate a new update cycle. The update state will reset to 0%, and processing will begin with the reduced batch size of 10,000 records per cycle. Records in the ml_cluster_detail table will be created as processing progresses. Expected Outcome After completing all three steps: the ml_solution update state resets to 0% and progresses normally, the “Error while updating solution” state is cleared, records are created in the ml_cluster_detail table, and the ML Readiness page warning is resolved.Related LinksTroubleshooting guide for Service Mapping ML Connection Suggestions / Traffic Based Discovery: https://support.servicenow.com/kb?id=kb_article_view&sysparm_article=KB0963421