By accepting the email your mail server has taken over the responsibility to handle it. If one of your downstream server then denies receiving it (and taking over the responsibility), your server has a problem and will re-try the delivery until either it succeeds or times out. As your server does not know, if this is a permanent configuration issue on the downstream server or only a temporary error, generally that behavior to retry make sense and as such is the default.
SMTP defined different for soft (temporary) and hard (permanent) failures, which might affect this behavior. Check your log files to find out the exact error code.
Basically you should re-configure your mail servers, so that have stricter policies at your front-servers and only relax them for backend-servers. Otherwise your front-end servers will always be in the (bad) situation, that they have to generate bounce-mails for mail, which cannot be forwarded to downstream servers and have to delete them, which results in data loss.
Have a look at the complete list of options man:postconf(5), but UCS only allows to configure a limited subset of that, e.g. body_checks_size_limit and UCRV mail/messagesizelimit.
