CMDB Health troubleshooting: failure threshold reached

The CMDB dashboard displays the string ‘failure threshold reached' when the number of CIs that are failing the metric tests, reaches the failure threshold set for the metric.

CMDB Health stops processing for this metric in the current cycle, and therefore there is no aggregated health score for the metric. Processing will be attempted again in the next cycle. Also, status in the CMDB Health Metric Status [cmdb_health_metric_status] table is set to Max Failures for this metric.

When the health score of a metric cannot be evaluated, then the processing status of the respective KPI (for example, correctness) is set to Incomplete. The CMDB dashboard displays the string Incomplete score for the respective KPI and for the CMDB Health overall score. Also, aggregated health scores for the metric are not available for any class in the CMDB hierarchy.

To troubleshoot, do any of the following:
  • Review and refine the rules defined for the metric which has reached max failures. If a rule associated with the metric is too generic, resulting in large number of failures, attempt to refine it. For Example:
    • Completeness – Review the recommended fields that are causing failures and remove the ones that are not critical for the health score. For more information see Set a CI field to be recommended.
    • Correctness – For the staleness metric, depending on the cause of the failures it might be helpful to increase the value of the CMDB Health system property.
  • Reduce the number of failures by fixing CI records: If after adjusting the metric test rules the max failures for some metrics is still reached, then address the failures by updating CI records with the relevant missing information.
  • Increase the failure threshold for the metric that is failing and check if processing for this metric completes successfully in the next cycle. Increasing the failure threshold beyond 500K might reduce overall performance.