Hello Niall,
I've longed submitted fault injections should be native to the cloud providers. It is fantastic that AWS shows the way with its FIS service.
However, not everything is lost with GCP or Azure.
Obviously, these are somewhat limited examples but it shows they are capable of designing for experimenting failures.
Additionnaly, one should never forget that all the cloud providers expose the API that can serve for injecting turbulence already. It's called "DELETE" :)
You can do so much already with calling directly the native REST API to delete resources. It's the entire ecosystem that has been built with the Chaos Toolkit (
https://docs.chaostoolkit.org/.
Sure it's not "Chaos as a service" by the cloud providers, but you can get very far with this approach.
Take care,
- Sylvain Hellegouarch