Runtime model plays big role in my case. As I frequently deploying different versions each time it will look for specific different blob paths (due to how internally different runtime model uses different locations to save it's status files) so this creates issue when one of the version is deployed other version's file stay as it is and after certain period of time if I switch back to that version and it status file is old and time has passed it triggers function.

So, yes, you can take note than if you are switching from one model to another please look for status file.