Plane crash in auto mode, need help to analyse log

I am flying an aerial mapping mission with my Skywalker 1900. I had flown two mission on that day, the plane crash on my third mission while it was almost finish the mission. It was a little windy at that time and rain was approaching. While in auto mode and flying along the grid line nicely all the sudden it turn to the right and start spiral drive and lost altitude rapidly and crash eventually. It was still somewhere in the thick tropical forest right now.
I have had great success with that plane and APM all this while. With my limited log analysis knowledge and skill, I can’t see any abnormal behaver of those sensors, the hardware voltage seems normal.
I am really asking for help to identify the problem before I make any adjustment or changes to the plane and resume mission flying. Attach herewith the tlog. Crash happen around 93.5 %

Thanks for your help.

Detail of my plane:
Skywalker 2014/1900
APM 2.6 external GPS and compass/power module supply by Jdrone
4S 5000 mah/ T motor 2814-710KV/12x6 APC
2.78 firmware

(Disclaimer: This is just my opinion and may be completely incorrect)

Your roll servo output goes a bit haywire at about 07h10m30s, both offset (from ~1450 to ~1790) and fluctuating a lot, plane struggles but seems to stay level with some wing rocking. Roll and Nav_roll also start getting out of shape about this time.

At 07h13 the roll attitude goes way out (from 0 to ~70°) but the servo output is maxed so cannot stop it, that’s the right roll you observed. The Nav_roll (requested roll) also oscillates wildly here (caused by flapping?).

It’s hard to say but it seems to point to hardware failure, perhaps one aileron came loose or the servo horn came off or a pushrod disconnected? Eventually the aileron broke or came off completely and the other wasn’t able to correct it causing the crash.

Perhaps a wing elastic broke?

The turn at 07h09 seemed very abrupt and the problem started after that.

Thanks Graham for your help. I also have a feeling of aileron failure. It will be good if it is true so I don’t have to worry about APM or the latest firmware. I can always build another plane but this time I’ll make sure all the linkage will be secure.
Once again, thanks Graham

Same diagnosis here. Attached graph shows ratio of servo output to X gyro diverging. Only possible answer is that a control surface, servo or wing failed.

Also of note is very high vibration levels (probably contributing factor) and .25v Vcc noise.

Thank you very much for the analysis. Really appreciate all yours effort. With a few flight again with APM, my confident regain. Heading for another mapping mission next week. Hope all will be well.