I have a list of some ~2500 logins. I’m running a script which makes queries to pull information about those logins. Essentially, it’s the same query over and over, accepting a login parameter. Each query “costs” one (1) point.
AFAIK there’s no way to “batch” these queries reasonably (which makes me wonder if a users endpoint is a feature that GitHub would consider implementing). I found that running these queries in series is very slow), so I was trying to run them concurrently.
Running eight (8) requests concurrently seems to trigger the anti-abuse mechanism, and temporarily bans me. Running two (2) concurrently seems to be safe, but it’s still much slower than I’d like.
This doesn’t seem to be documented anywhere, so I’m hoping GH can give more clarity around this (unless it’s risky to be more specific). What’s reasonable?