Spot instances keep being put into the dirty “FinOps only” corner, even by prominent GreenOps players out there. This talks aims to get rid of that image and explain how spot instances can, in many ways, profit sustainability!
First, we’ll take a look at the bigger picture: what are spot instances, anyway? Why are they a GreenOps tool, although a slightly tricky one? What does the energy transition require from IT systems and how do spot instances support that?
Then we’ll talk about how to actually make spot instances work in a project without totally ruining your ops landscape, like they often do if you’re not careful. I’ll bring two examples:
- one where a team had huge problems with them, explaining what happened, what caused those problems and how their use had to be rolled back. And
- one where spot instances usage was automated using Open Source tool Karpenter, showing how to do that in your project and how our customer hugely profited for over a year already.