Hey there after struggling a bit because of my upgrade from 9.1.2 (problems related with the system.url-prefix which i had to change). I ended up with one host with Nginx (to deal with the acme.sh for SSL ) and 24 containers , where one of them is again an Nginx a reverse proxying serving some upstreams to other containers.
Is it too stupid from my side to make some changes to that container so i can add the acme.sh directly on it ?
Are you planning to maintain this container as a reverse proxy ?
I was thinking in adding the repo from acme.sh , make some changes on the nginx.conf , so i could accept the 443 requests and some environment variables to pass the domain to request the SSL.
It’s not stupid but I’d definitely recommend against it.
Yes as this is now required by the new Relay service that became the event ingestor (compared to sentry web in earlier versions)
I’d just use a separate, external Nginx/whatever service for this. It may feel wasteful but the whole Sentry on-premise package should be treated as a black-box as we may need to change the configuration or nginx itself in the future and then you’d have a hard time migrating.
Thanks for the explanation . It makes sense 
1 Like