AAISP seem to have a number of systems that are single points of failure. While they are quick to respond to outages it would be nice to see some resilience in the server estate to reduce the number of potential outages.
there are ongoing projects to improve resilience in all areas. Since this post was first made (2011) there would have been many changes in how services are provisioned.