Sometimes, subprocesses only log a warning if the return code is non-zero. For example, running a fresh installation of Fedora 38 and build #3414, I get a zero exit status from sendmail even if the process writes to stderr, and nothing is written to the log. This can make errors difficult to detect. A simple fix would be to also check if stderr is empty or not as well whenever we run a subprocess. stderr from sendmail (return code 0): > No configuration file found at (null) or /etc/esmtprc