We experienced memory buildup and service crashes because of exception storms in our dockerized services. I know now this is due to a change in version 10 setting GSS session encryption ssl mode to 'prefer'. We found that in a high traffic environment, we got around 50k exceptions of 'Exception while performing GSS encryption' in the first minute, causing a memory buildup and cpu spikes. This happened in all of our services more or less (APIs and worker services). Disabling it via the connection string makes the problem disappear, however, I wouln't say 'the error is harmless', as stated in the docs.

Now 2 things could be going on here if there are no GSS libraries present in the container:

1. the error gets thrown internally repeatedly
2. we configured something wrong with npgsql and something gets reinitialized

We've been using npgsql (and with ef core) for years in prod without issues:

            services
                .AddDbContext<SoiDb>(options =>
                {
                    options.ConfigureWarnings(w => w.Ignore(CoreEventId.RowLimitingOperationWithoutOrderByWarning));
                    options.UseNpgsql(Configuration.GetConnectionString("db"), o =>
                    {
                        o.UseNetTopologySuite().MigrationsAssembly(migrationsAssembly);
                        o.MigrationsHistoryTable(HistoryRepository.DefaultTableName, "ourdb");

                    }).UseSnakeCaseNamingConvention();
);

i couldn't find anyone posting an issue about this so this makes me think it is something we configured wrong?