Internal DLR storage is not the ideal on a production environment.
Upon a Kannel restart or crash you will lose all your DLR information...
Using a DB for their storage is much better.
Depending on the traffic you got, the DLR table in database might grow
considerably too, but if you need to
Today one of our kannel instances became increasingly slow. After some
troubleshooting, we realized that the internal DLR storage was too big and
the threads were taking too much time to store a new DLR.
Is there some configuration to cause expiration of the DLRs in memory?
I could not find