First, thank you for your quick reply.
Since you list two windows versions, is the new Lucee 6 running on both? And does the problem seem the same on both?
Lucee 6 is running on both Windows Server versions (2016 and 2019) and the problem is the same on both.
The only difference is that the server with Windows Server 2016 has more than 100 datasources than the one with Windows Server 2019 (only 2).
Or might the old machine run the old Lucee version? If so, have you tried running Lucee 6 on the old machine? Even just for a test (if not prod traffic), to confirm whether the problem happens with Lucee 6 on the old machine? If it does not, that exonerates Lucee and points to some other change in the new box (or network, security, a/v, firewall, etc.).
Machines are the same, only Lucee version was changed from 5 to 6. Everything else is the same.
Along the same lines, is the sql server unchanged between the Lucee 5 and 6 calling into it? Or might each point to a different sql server, perhaps as different versions or each on different machines themselves? That could be important to know.
SQL Server is unchanged, both machines are using Microsoft SQL Server 2022.
But we have a third machine with Lucee 6 and also Microsoft SQL Server 2022 and it is working fine. Difference is that it has a datasource connection limit (max) of infinitive, but in my opinion that should not be infinite.
And do you have any dsn connection string values out of the norm? (And if you may define the DSN in your app to override the same one in the admin, be sure to assess the former.)
We are not overriding anything regarding the DSN in our app. Datasources are defined in Lucee admin only. Some datasources are created in admin also using cfadmin api, but with the same settings as in Lucee admin.
How are you measuring the connections? And are you observing things from both the Lucee and db side?
Finally, are you using anything to measure the rate of traffic (into Lucee, or into the db), confirming there is no dramatic increase to explain what you have observed? Don’t presume “it can’t be that” as “nothing changed but the Lucee version”. If you do measure the traffic, let us know what you’re using. It could matter.
No, we are just limiting datasource settings using:
- Connection limit (max): 200
- Connection idle timeout: 10 minutes
- Connection timeout: 15 minutes
Currently we are not measuring it but only monitoring it using Fusion Reactor and SQL queries in SQL Server. When our app slowed down we checked datasources in Lucee admin and we saw a number of 200 in “Active” column for problematic datasource.
And customers started complaining about slow response times and timeouts. 
Nothing more.
Even if someone else may have a different “answer” for you, based solely on what you’d offered, if you offer brief replies to the above it may help refine what they’d offer."
Also saw that other people have similar issues: