Yes! DB-less mode will be an option that can be turned on. If you don't want to use DB-less, or your configuration is too large to be held in-memory, you can continue using a DB.
To date, Kong's configuration has lived in a DB, but with DB-less you can hold the entire configuration in a single YAML or JSON file. You'd want to use DB-less in instances where you want to use declarative config, or in instances where you want to minimize the number of moving parts. As long as your config can fit in-memory, it's a great way to keep things simple.
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.
8 comments
[ 0.83 ms ] story [ 24.6 ms ] thread