Hello, Marco CTO of Kong. You can choose between two modes for Kong now. Kong with database (PostgreSQL or Cassandra) so that the configuration lives in the database and can be propagated automatically across any Kong node that connects to the same database. Or, starting from v1.1, Kong with no database dependency (db-less) with configuration stored in the declarative configuration file which now becomes the source of truth. In the latter, as long as every Kong node starts with the same declarative configuration file, every Kong node is effectively looking at the same data.