Christopher Gutteridges’ talk was full of practical tips to manage the workload of web services management, making you more like Batman with his Batcomputer than Superman rushing to the rescue when things suddenly go wrong.
Chris argued that it is a waste of time and resources to have humans doing jobs that computers can do really well. To this end, his department has written lots of scripts to perform basic tasks – particularly monitoring tasks – to help make them more visibly more efficient.
Whilst monitoring their department’s systems in this way has made it easier to manage their workload – preventing them from dropping the ball on normal requests in favour of crashing in to the rescue when something unexpectedly breaks – Chris had also noted that this practice was going to be especially important in the current economic climate, when we are all facing uncertainty about our jobs.
Monitoring is enabling them to build a stock of statistics and graphs that show how much work they do for the department they support, and how efficiently they work. He was concerned by the amount of data people in similar roles in other institutions could be throwing away and argued strongly that they should start preserving that data – even if they don’t do anything analytical with it initially. He also recommended monitoring the things beyond your control that cause your systems to become unavailable.
Throughout, Chris kept returning to the point that we are here to facilitate research and teaching. Building a Batcomputer can help you focus on this task and allow you to better enable your researchers and students, rather than rushing round trying to be Superman.