bigfalcon wrote on Wednesday, December 12, 2018:
Hi! First time posting here. Hopefully this is not a stupid question.
I come mostly from a mobile app development background. But recently looking into a project that involves essnetially a Cortex M, currently prototyping on an Arduino, and considering using FreeRTOS.
I am very used to remote crash reporting solutions like Sentry or Crashlytics for mobile development. I am wondering about the same thing for embedded devices. How do one ensure an anomaly free experience once the device gets into the end users’ hands? Or is this somehow not a realistic concern?
Please advise. Thanks in advance!
rtel wrote on Wednesday, December 12, 2018:
I think the reporting is somewhat more simplistic on a Cortex-M,
especially as the RTOS kernel and application are built as a single
executable, rather than there being the RTOS running separately from the
application. Some tools, such as Atollic, have some minimal crash
analysis built into the tool (type of exception, address off offending
instruction, register values). If you don’t have such a tools then you
can add some basic code as per the following link to help:
bigfalcon wrote on Thursday, December 13, 2018:
Thanks for the tips on Atollic! Yeah I figure we can build a custom tool for field reporting. But we are wondering if there are any turn key solutions or common design patterns around this. Or is this simply not a common problem - detecting issues that end-users might be having in the field without adding user friction in doing error reporting?
We at Percepio are working on a solution for error reporting in deployed devices. This will be released in mid February and is initially targeting FreeRTOS and AWS. We are running a beta program with early adopters, so if you are interested, just let me know.