Methods like Discovery, ACC-V or Service graph connectors (for integrations) are required for automatically populating relationships. Using these methods also, the number of relationships populated will depend on whether these methods support the population of relationships for the CI types present in your organization. If you don't use these methods, you need to load the relationships manually into cmdb_rel_ci table.
Also consider the following points:
1. CMDB should track only those CI classes and attributes which support some business objective (E.g. IT service management or IT asset management). Avoid populating any CI records which does not serve any purpose.
2. If you actually need to track a million CIs, automation is definitely recommended. Use of only manual methods will definitely lead to data quality issues like duplicate CIs, stale CIs etc. for such a huge volume of records. Consider using Discovery or Service graph connectors (if you have existing tools like SCCM or JamF to integrate
3. It is important to constantly monitor the CMDB data quality and take steps to improve it. Make use of 'CMDB health dashboard' and 'CMDB and CSDM data foundations dashboards' for this.
