We have had this issue with an application of ours as well. I believe there is currently a bug in how multi AZ is being handled since the upgrade of CF buildpack in June.
I would contact support regarding this if I were you. As far as I know there no public statement regarding this issue and potential solutions as of now.
@Bharathidasan as Joris describes the performance problem dissapears the moment multi-AZ was disabled on their production environment.
Both before and after the multi-AZ was enabled the DB was running circles around the application instance(s) and there was no obvious part of the application that slowed down the process. Random retrieves took 30+ seconds to process. An application restart helped a bit but not for long.
It really seems that multi-AZ has one or multiple flaws contributing to the issue. Such as perhaps having the DB running in a different environment then the application instance(s). It would be great if Mendix could elaborate on this.
The possible reasons for this issue could be related to how data is stored and transferred between different parts of the app, how much computer power (CPU) is used, and how the app is designed.
To fix the problem, you may need to check how much computer power your app is using and optimize how data is stored and retrieved.