Closed JohannesEbke closed 5 years ago
The reason for this is that AWS has over 1000 endpoints (region/service pairs), and aws-list-all distributes requests randomly across these endpoints to avoid throttling.
Workaround: If you list regions one-by-one, you should not run out of open files.
If run with a nlimit of 1024, aws-list-all runs into OSErrors due to too many open files (connections).