Tweak timeouts for scalability

[geonnave]

At the large deployment in Limerick in the end of January, we observed:

1. 600 bots: all worked very well (only a few bots failed flashing)
2. 725 bots: half bots were failing flashing, at the time I didn't find a solution, but now I think that tweaking timeout and retries could address the issue

Given that, I think it could make sense to have a profile for the timeouts/retries, based on number of connected bots. For example, those values could vary based on whether there are up to 10, 100, 500, or 1000 connected bots.