Camunda engine architecture is essentially an event sourcing architecture and it uses a database to essentially manage a materialized view of the engines internal state. This db is not intended to be for general use…
You can connect camunda to an external db using one of a number of patterns depending on use case…
You can stream historical events from engine to an external db…
If a task requires access to an external db, you can use code to connect to a db using a task worker. Another low code option is to use a connector…