To create a new iNZight session, you need to be a person. I have used the reCaptcha tool to do this -this requires a person to enter numbers of letters using visual or aural cues.
Once the docker instance is activated, we wait 10 or so seconds for the web server to load the updated configuration… After the new container is activated, the nginx server reloads the new configuration file associated with the proxy used to make the instance accessible on the host’s web service. 10 seconds is probably too long, but seems to work most of the time.
…and from here you can connect to the new instance -remember that address, it is unique! The address generated is specific to the session, and will not work after the instance has been stopped and deleted. The chances of another instance getting the same combination of alphanumerics is slim. This makes it hard to remember or share, but I believe this is necessary to protect user data from people who are not part of any group activity.
I am pondering using randomly generated easier to remember names (as docker itself does).
Now the iNZight application is presented – shiny-server in action! From here on in, you are dealing with iNZight -a software environment developed in Statistics.
Play with the data!
Multiple users can access any one session, but the idea is that each of these is private to a group at a specific time. Idle instances are stopped after 30 minutes of inactivity -measured through web server activity. You have browse away from the iNZight pages, shiny-server maintains a constant flow between client and server, so simply leaving the web site open should be enough to keep the instance alive. Stopped instances are deleted every 15 minutes or so.