We did file the initial bug report and we believe we found the cause and remedy for this bug: When deploying the Infra L3Out for B2B, we accidentally created two seperate Fabric External Routing Profiles (under Tenants -> Infra -> Policies -> Fabric Ext Connection Policies).
After deleting the duplicate entry under Fabric External Routing Profiles, the OSPF crashes stopped occuring. Usually they happened in a four to six week interval, after removal of the duplicate entry, we haven't had a crash in ten weeks. So we believe this fixed it. Just in case this is useful to some other customer or Cisco's dev team.
