The application is a single page, so it is not updated, and we hope that it will work on the Android device for a long time.
We are currently seeing a crash (and a logcat report indicating an OOM error on startup on Android) after about 18 hours. When I use remote debugging and fix the timeline using the chrome dev tools, it is very obvious that we have a memory leak, because we see a steadily growing bunch of js. It seems like the heap continues to grow until I force the garbage collection.
Another thing that seems special to me is that on the timeline, I also see that our event listeners are growing and not affected by the forced GC. Could this be a problem in itself? (We are talking about 10 thousand listeners according to the timeline)
Does anyone have any recommendations that go beyond the standard “3-way snapshot analysis methods”? This is not very useful, because the heap grows even when it does not interact, and I believe that this is due to some periodic updates that we do behind the scenes to read and display data. Otherwise, I am completely open to any other memory analysis tools.
I have very little experience in this area, so any advice on narrowing down the cause of these unpleasant leaks will be awesome!
, - . .
EDIT:
, , dev tools , , ... ?