I'm having trouble finding the origin of a memory leak I have in my node.js application.
I'm using a HTTP server running on each worker of my cluster application. Each worker handles file download & upload and I can notice a huge memory leak (that can go to 6 to 8 GB in res memory).
I tried getting a heap snapshot of one of the worker leaking heavily and having trouble understanding it.
In the beginning of the dropdown list when selecting the first and heaviest JSArrayBufferData we can see that there's a Buffer that holds all the memory, but if I scroll down enough the Retained Size is also listed under " in Socket".
I don't really understand from the Snapshot if there are some socket in my application that retains file informations and causes the leak or if a timeout is causing huge trouble.
Any help understanding or pointing me toward more information about Heap Snapshot analysis would be greatly appreciated.