Mobile UEM help
Hey everyone, my customer has some questions around Mobile UEM. I think I know a few but could really use some input:
- Do you support installing from Carthage (vs CocoaPods)?
- I'm pretty sure this is no.
- Are there issues that arise if we are using obsfucation and various checksum guards as part of runtime and build process?
- In this I think they are referring to their class and method names
- How big is the SDK, what performance impact, how much RAM is used?
- What can/should we do to instrument the app with code beyond the base install?
- This depends on the app right?
- How do we handle recording the length a time a function takes when it is a multistep process (for example. Remote start: create request (get id), poll id for status until complete)
- To what extent are crash logs recorded (stack trace?) – compare to apteligent
- We just capture basic crash details I thought
- What data is recorded from the device (we need specific fields for legal/privacy review). Can we stop certain attributes from being recorded?
- By default its just IP address I thought.
- Is there support for recording code generated events to track, such as an exception?
- Not sure what we capture in mobile UEM for this.
Thank you in advance for the help guys.
1: no carthage support
2: Android: the mobile agent library classes and webrequest library classes (e.g. apache, okhttp) must not be obfuscated otherwise they cannot be instrumented
iOS: runtime checksums might cause troubles for methods swizzled be the mobile agent
3: that question cannot be answered in general as it heavily depends on the architecture o the app (how many viewControllers/Activities, how many web requests). Android instrumentation happens at build time so there is no instrumentation overhead during runtime, just the additional code added by instrumentation and agent communication sending the data. For iOS instrumentation happens on Application startup therefore most noticeable agent impact is during App startup and heavily depends on amount of viewControllers that need to be instrumented. Rest of runtime overhead is similar to Android
In general max agent runtime overhead is communicated with 3%
4: depends on the app and what the customer wants to se
5: this can be done with manual instrumentation by defining the start end and of the process
6: Android: stacktrace,
iOS full crash report (including stacktraces of running threads, loaded libraries, device information
7: IP adress is recorded on webserver agent/java agent that receives the mobile beacon - this is not directly recorded by the mobile agent (I'm not sure if that can be prohibited)
The mobile agent reports ist own version, OS version, device Manufacturer, device model, app version, RAM size, free RAM, CPU type, screen size, orientation, battery level, and carrier. Those cannot be turned off. Only GPS location reporting can be turned off, but GPS is only captured by teh mobile agent if already captured by the mobile app.
8: you can track everything with manual instrumentation. it allows to create actions (duration) or events (string and other basic values) or errors that are reported to the AppMon server
Hope this helps,
Patrick
