I'm wondering what are the potential problems I might face if I override the Kubernetes metadata name in my declarations? This means my Kubernetes resources will get the exact name I provide and not the one which is the resource name plus generated string? There's several reasons why I would want to do it but I wonder if this will lead me to some kind of world of hurt?

The main thing you’ll need to be aware of is that any resource you create this way that cannot be updated in place will be deleted before it’s recreated if you make changes to your infrastructure. I usually only do this on namespaces

https://www.pulumi.com/docs/intro/concepts/resources/#autonaming gives some more context on this

Thanks @billowy-army-68599, this was what I was suspecting. In our case downtime is not an issue and avoiding unnecessary resource scaling (due to temporary duplicate services) is more important so this is actually exactly how we want it to work 👍