Mistakes happen and occasionally a typo when provisioning nginx/sslterminator roles can cause the service restart to fail and the live sites to go down as well. At that moment the stress level goes goes up exponentially until you figure out the problem and re-provision the roles.
We could implement a nginx configuration test to be done before service restart. So nginx -t check and if it was successful allow for the restart otherwise just fail provisioning. Site(s) will stay up and the provisioner can fix the error and try again.
Mistakes happen and occasionally a typo when provisioning nginx/sslterminator roles can cause the service restart to fail and the live sites to go down as well. At that moment the stress level goes goes up exponentially until you figure out the problem and re-provision the roles.
We could implement a nginx configuration test to be done before service restart. So
nginx -t
check and if it was successful allow for the restart otherwise just fail provisioning. Site(s) will stay up and the provisioner can fix the error and try again.