I am shifting my database to Aurora Cluster MySQL. For testing, I have created a small sample data (2 GB DB size). I am connecting to the cluster writer endpoint. Connections and queries work fine for a while but after 10-12 hours of running the DB, few queries start timing out (of course these are queries which take a bit more time).
I have modified the parameter group to increase net_read_timeout=6000 (bit too much) and also a few other related parameters which ensure connections are not closed. Even the load on the server is not much (so number of connections is very low).
But after every iteration, when I reboot the database, it works fine for 10-12 hours but connections start closing after that.
The issues are multiple like
Lost connection to MySQL server during query or
Malformed Packet etc.
One thing that I am suspicious about is that it is a db.t2.small instance.