Post-mortem: The certbot package defined a systemd timer and a /etc/cron.d/ entry for renewal, so my own crontab entry that defined a post-hook never got to do the renewing
@benofbrown I write a script call by cron to check if the NGinx was still alive, after 3 try (10 minutes total) it try to restart the service. @Gargron
@C_Chell@gargron Restart or reload? Restart's probably overkill.
It works fine now with my actual hook in place. I noticed it before the cert expired as my hook emails a notification and that notification wasn't being sent.
@benofbrown On my configuration, I notice NGinx was reload but at the same time, certbot didn't release the ports so NGinx won't load the ports and I need to do a "service restart" to force NGinx to relisten on 80 and 443. @Gargron
@benofbrown I change it recently, but I have to recheck the configuraiton soon to be sure.