After FACS sorting and prior to loading cells onto the 10X Chromium machine, one or preferably 2 wash steps (centrifuge @500g, 5 mins) are required to reduce contaminants and ambient RNA (released from dead/dying cells). The cell density should then be quantified using a cell counter such that an appropriate number of cells can be loaded.
The “optimal” situation according to the 10X user guide would be to FACS sort 100,000 cells, wash twice, and load 7,000 cells. However, this is often not possible due to limited cell numbers. In this case, FACS sorting 50,000 cells and washing once should allow the loading of 7,000 cells. The disadvantage of washing the cells only once is that this might lead to a higher amount of ambient RNA captured in the 10X droplets, but if this is the case there are computational tools that can detect and subtract ambient RNA (e.g. SoupX) from your samples during analysis and I have had success with this approach.
In my own experiments, using 120-150 whole larvae (3dpf-5dpf) I sorted 60,000 cells (50,000 macrophages +10,000 keratinocytes) and was able to comfortably load 7,000 cells. I think you can get more macrophages out than that (possibly 70,000), the keratinocytes were the most limiting factor for me