I have a .net-6 project that periodically downloads archives, parses xml data from them, converts it into classes and then stores them as entities to Postgres via EntityFrameworkCore.
The data being inserted is quite bulk (every archive turns to thousands of entities) and takes time and memory.
When I launch my project with memory profiling via Rider, a memory "saw" has quite big "teeth" but stays stable, on every GC-tion memory always falls down to an average level which doesnt grow.
When I deploy my project as a kubernetes pod in our corporate cloud, that average memory level constantly grows so finally I get an OOM-exception.
Please, any ideas, why this can happen.
I tried enlarging memory limit, expecting the average level finally reach a "shelf", but it didn't work.