One of TJ's computing clusters (40 nodes)
Recently, based on average data, it looks like this system has gone down for about 5760 minutes at a time.