My tests show that iRedMail appears to whitelist a server after it passes the greylist test. Would like to have a better understanding of the implementation of greylisting in iRedMail and be sure of the conclusion since the test was a small sample.

Additionally Gmail or Hotmail for example may send from many different servers (server1.hotmail.com, server2.hotmail.com). Is the greylisting domain based/ip based/ ect?

The only issue I have found with greylisting is that, very occasionally, mail can get stuck in loop if the sender domain uses multiple sending servers on different IP's. If the sender has a round-robin approach to the sending server then the greylisting refuses delivery of the 1st; 2nd; 3rd IP's and by the time the 1st server tries again the greylisting refuses again due to the length of time since the first try.
This issue has only been very occasional - it also affect the Barracuda hosted spam service we use for work email too.

ZhangHuangbin - if you have a backup MX specified that doesn't use greylisting but your primary server does use greylisting - is there a simple way to stop connecting servers from immediately delivering to the backup mx when the primary mx refuses during the greylisting period?

Thanks for a great mail server btw any thoughts on a docker app?

It sounds like a sender-side issue, not yours. no idea.

We're lack of man-power. If you're interesting in build a Docker app, i'm willing to offer help.