As an admin I have configuration recommendations for low-bandwidth scenarios

[ipanova]

Describe the solution you'd like
As an admin I have docs which help me understand what values/settings to tune to overcome bottlenecks during container image push

Depending on the user base there might be quite a lot of concurrent image push also push of large images.
Admins need to know how and whether it is possible to improve the situation whenever 503/504 errors are surfaces to the end user.

There are server/system config that can be tuned like nginx values, socket max connections, number of file descriptions, worker timeouts, etc.

• client_max_body_size, the default is 1m, it can/should be adjusted for a more performant podman push
• gunicorn timeout can/should be increased to mitigate 504 error during a lot of simultaneous podman push which would lead to i/o saturation.

Network bandwith, disk iops, memory and cpu values need to be taken into account too.

Additional context
https://issues.redhat.com/browse/AAP-1729