Because of the time involved in generating static JSON data and the the delay in syncing that data between different machines, the data displayed on the main page of slave health can sometimes be stale by as much as 30 minutes. This isn't helpful when sheriffs and buildduty are trying to make assessments about the current state of the tree. They of ten need to go check other data sources for timely data. While it shouldn't be the default, there needs to be a way for interested parties to get essentially live pool data from slave health. We could implement this in different ways: * generate the static data more frequently * use static data to build the initial page, and then lookup fresh data async once the page is loaded. * allow the user to click on a given slavetype to get fresh data for just slavetype * ...?

slave health is retiring with Buildbot in August.