IOS app crashes in managed access mode

Question

I am working on a team that developed an application whose typical use case is unlimited continuous maintenance of managed access. In our latest release, we came across one client who repeatedly came across instances of their devices, which seemed to fall into our "emergency" mode on the main screen. They stated that, despite the fact that they were on the “home” screen, the devices (we were told that the client was not in place) were still in “Guide Access”, which made no sense to us while we ourselves We didn’t test the script, performing a special assembly that would lead to a failure and monitoring of the results. We discovered:

• after 5 crashes of our application in managed access mode (and “Guided Access” restarts the application each time it was developed), “Managed Access” will then basically fail and display the keyboard to enter the access code to leave “Managed Access”.
• if we didn’t interact with the keyboard after some time (about 1 minute), the keyboard will disappear and the home screen will be displayed, but the user will not be able to interact with the view, except for pagination pagination to view the second page of application icons. Touching the application to launch it will cause some animation to recognize the touch, but the affected application will not actually start. The device was then basically in a brick state requiring reset.

We are testing in iOS 8.x.

I researched online, including the Apple Developer Forums and the iOS documentation, and have not yet found an explanation of how often 5 crashes happen (if it's always 5) so that Guided Access basically says “no mas” and ask the user to log out from managed access.

So my question is: is there a certain speed of crashes that must occur before Guided Access gives up trying to restart the application? Is it 5 outages in 24 hours, a month, or 5 outages no matter how long it takes to achieve this in a single managed access session? Our application is usually used for kiosk applications, so the application may be unlimited. We confirmed that 5 failures within 4 hours would actually create the scenario described above, but instead of testing days / weeks to get an answer, we decided that we would turn to the SO community.

In addition, I understand that the true solution to the problem is to fix our code base in order to prevent a crash - just trying to understand how vulnerable we are to this scenario.

Thanks in advance.