No, this is not normal. But compass errors are usual reason for fly-aways. This is a SW thing when things happen that the programmer did not expect. The flight controller SW cannot handle the wrong sensor data in some cases.
From my point of view this is a warranty case. Flight log file is there as evidence.
Timeline:
Take off: 01:10 after boot; Start of compass errors (geomagnetic field stuff): 01:26; Critical positioning warnings 01:38/01:39.
I have compared stick input and the behaviour of the copter as result.
1) Throttle: It follows the stick input until 02:30 after boot. Then it climbed by itself.
View attachment 18346
2) Pitch: It follows the stick input all the time. Don't forget, you are in manual mode since 01:39 (29 sec after take off).
View attachment 18344
3) Roll: It follows the stick input until there is an unexpected spike in roll at 02:28.
View attachment 18345
4) Yaw: It follows the stick input all the time.
View attachment 18347
This means, you could control the copter at least until 02:28 (78sec after take off) but it was too far away to see how it reacting.
View attachment 18348View attachment 18343
br HE