Offering Database As A Service can be one of the biggest challenges when working with Kubernetes. Maintaining, updating, and collecting metrics from not one, not ten, but thousands of databases, when many of them live in their own, separate Kubernetes clusters would be incredibly toil-heavy, without proper automation being in place.
Unless you are working for a mega company like Microsoft or Apple, you're likely to be limited by the number of engineers who can take care of the product, meaning that you have to discover smart approaches to achieve a high level of resilience and scalability.
This talk will cover different approaches we have taken to ensure that our databases are secure, reliable, and that we can roll out an update on all of our databases within a couple of minutes without causing any downtimes to our customers.
We will look at some of the technical implementations we've put in place to ensure that we are capable of managing thousands of databases with only a couple of dozens of engineers. We will talk about decentralised architecture in Deployments, Incident Management, and we're also gonna have a quick look at some of the SLOs that we use to maintain a high level of resilience.