`systemctl reload teleport` sometimes hangs with defunct process

[one000mph]

Several times after updating my config file I run systemctl reload teleport and am seeing the process hang

$ ps aux | grep teleport
root     12755  1.3  1.5 624452 59800 ?        Ssl  09:46   0:06 /usr/local/bin/teleport start --config=/etc/teleport.yaml --pid-file=/var/run/teleport.pid
root     12869  0.1  0.0      0     0 ?        Z    09:54   0:00 [teleport] <defunct>

In order to resolve it I kill the defunct child and parent sequentially. After that I am able to start up teleport with my new config.

I'm not sure if this a bug, or an expected result of restarting the auth or proxy service. The times I experienced this I didn't have any open connections to nodes or to the proxy, but I'm sure there are some intra-node networking happening which might be hanging the graceful reload.

Hopefully this report helps with future development since it is an unexpected usability issue.