Following are some best practices to consider at a high level for IRE:
1. Service Graph Connectors (if available) should be used for 3rd party tool integrations to populate CMDB. This will follow IRE OOB without any additional customisations.
2. Integration hub ETL should be used for data imports to populate CMDB. This goes through IRE. Use of normal data import and transform will not consider the IRE.
3. For all the key CI classes that are used, review the identification rules currently configured. Based on the unique attributes of the data sources which are populating each class, the identification rules should be refined if required. This should be done on priority for classes which are having too many duplicate records. Use the metrics like duplicate CIs from CMDB health dashboard and 'CIs Processed via IRE' from CMDB data foundations dashboard to identify the CI classes for which identification rules might have to be refined. Also for each CI class populated by more than one data source, data precedence rules should be defined.
