SREcon18 EMEA

There have been computer emergencies as long as there have been computers and emergencies. There is evidence dating computer on-call shifts from the 1940s, and we can trace on-call activity in an essentially unbroken line from today back through seven decades of the computer industry.


But just because we've been doing it for seven decades doesn't make it right. In fact, if you think about it, the fact we've been doing something essentially unchanged for seven decades, given everything else that has changed around the practice, might cause us to ask: is it right that we are doing this? Is there something better? What are the alternatives?


This talks builds on previous work to analyse and establish what the _real_ reasons for continuing to do on-call are, provides evidence that human beings are actually really bad at it—in fact, it's really harmful for them to do it—and proposes solutions, ending with a call for action for the industry to bring into being an on-call-free future.


(Note: this talk builds on an article written for DNB's "Seeking SRE.")