[Check_MK] Monitor nginxplus config regularly

[acozine]

User story

As an operations person, I want to know when the inactive load balancer is failing its nginx config check, so I can fix it before the active load balancer bumps traffic over.

Acceptance criteria

• A check in CheckMK runs nginx -t regularly, knows what failure looks like, and alerts us.
• Test by pushing a config file with an error in it to the inactive load balancer and confirming that we get an alert to Slack and/or elsewhere saying that nginx -t has failed.

Implementation notes, if any

Start with the dev load balancers, then expand to production.