I cannot help you with a solution for the error because I haven't seen it before. However, I've had contact with Mendix support converning the number of DB connections in one of our apps. You can indeed raise the maximum number of connections by adding the ConnectionPoolingMaxActive & ConnectionPoolingMaxIdle settings. The default for these settings should indeed be 50. You'll also need to contact Mendix support beforehand because the cloud ops team will also need to add some configurations if I'm correct. If your and their configurations don't match, you could experience some undesired behaviour (Errors like: Opening JDBC connection to 10.140.51.87:5432 failed with SQLState: 53300 Error code: 0 Message: FATAL: too many connections for role "..." Retrying...(1/4) )
There might be something else going on. Are you using a reverse proxy between Mx and your end users? Because this could also cause the issue. Also have a look if your users are logged out immediately or if the experience a delay before they are logged out. If the delay is there make sure that you are not having long running queries open.
I had a look at the mee documentation and found that mendix can close the client connections as well.
# Abort database SELECT queries that are started from a client XPath request,
# or XLS/CSV Export button and run for a configurable amount of time.
# The reverse http proxy in use might have a proxy gateway timeout set (which
# is by default 60 seconds when using Nginx for example), so continuing while
# nobody can receive the results anymore is a bit pointless...
#
# Setting this option prevents runaway database queries from eating up all
# of your database cpu cycles, while you're busy tracing down the source of
# the problem (using LogMinDurationQuery, see below)
#
# This option was introduced in Mendix version 2.5.6
# The value is specified in seconds.
#
# default: not set, no timeout
ClientQueryTimeout: 70