aws-eks deployment readiness

[tomklapka]

• updated readme
• added aws-eks.sh deployment script
• changed create.sh script in root to be able to deploy in different providers
• added resource metadata to create necessary accounts and roles in kubernetes
• contains fix for /stats/summary returned HTTP status 401 Unauthorized #77

[jackzampolin]

This is a pretty substantial change. Ideally the create.sh script could run against any kubernetes API that the local kubectl/helm is configured for. Is there a reason to make this specifically to point at EKS?

[tomklapka]

This is definitely a valid note. There could be small but significant differences between different providers (e.g. different namespaces, no or different external access, roles etc...). After retesting it on other providers it's possible that I come up with something more general.

[russorat]

@tomklapka can you confirm that the original functionality still exists? A user should be able to spin up the entire stack on a locally running k8s by default if there is no provider given?

Also, is there a way to repurpose the create.sh and destroy so that they still work for now, they just call the run.sh script with a note at the top saying that the user should use the run.sh script in the future.

[tomklapka]

Update: re-unified run.sh script - it's now able to deploy all components to both minikube and aws-eks with minikube as default. Every provider needs small adjustments (already included in the script or mentioned in readme) to make everything work ootb.