Add health check location

[jmastr]

Ideally, this feature should be a part of a bigger feature (the default
server) that solves #52.

For now, we add the "-health-status" parameter to the controller. If
it is present, the default server listening on port 80 with the health
check location "/nginx-health" gets added to the main nginx
configuration.

Closes #90

[pleshakov]

@jmastr Thanks!

Any reason you chose the default type as default_type text/html;? For the response that NGINX is sending, text/plain suits better.

[jmastr]

@pleshakov It was for debugging only. I wanted to stop my browser downloading "healthy\n" all the time. I removed those lines.

[pleshakov]

@jmastr Could you add it again as the text/plain?

[jmastr]

@pleshakov Sure. Pushed it. But it is the default anyways, isn't it?

[pleshakov]

@jmastr Great! The default is default_type application/octet-stream;

[jmastr]

@pleshakov I am just curious, am I looking in the wrong place: http://nginx.org/en/docs/http/ngx_http_core_module.html#default_type?

[pleshakov]

@jmastr That's correct. Sorry for the confusion. However, this default is overwritten by this line https://github.com/nginxinc/kubernetes-ingress/blob/master/nginx-controller/nginx/nginx.conf.tmpl#L16, which becomes the default :)

[jmastr]

@pleshakov Thanks for the explanation. Everything looks good to me now.

[jmastr]

@pleshakov btw. I am just preparing my talk about Ingress and NGINX Ingress Controller for tonight's Kubernetes meetup. Told you about it: https://www.meetup.com/Berlin-Kubernetes-Meetup/events/235968196/

[pleshakov]

@jmastr Awesome! Good luck! :)

[pleshakov]

@jmastr How did your talk go?

[jmastr]

@pleshakov It went very well. The people were pretty interested. I could have talked a lot longer about that topic. There was one guy already, who manages 100+ vhosts with the NGINX Ingress Controller. Keep up the good work!

[pleshakov]

@jmastr Glad it went very well! thx!

[georgecrawford]

@jmastr I know this isn't yet ready for public consumption, but I did try it. I've tried adding -health-status as a command param to the container, and I see that the expected block is added to /etc/nginx/nginx.conf. But at this point, I don't see my custom settings as defined in -nginx-configmaps. However, moments, later, I see that the main config file is re-written, with my custom settings from -nginx-configmaps, but without the -health-status block. It only therefore persists if I don't set -nginx-configmaps at all. Do you agree this is a bug? Want a new ticket?

[jmastr]

@georgecrawford I did not test it with configmaps indeed, so the probability that it is a bug is quite high.