Introduce Instance Field in the Workload and App Spec

[thschue]

Synopsis:
To differentiate between multiple environments/instances/stages of an application, there needs to be something that helps distinguish them.

Therefore, I propose an introduction of an additional field in the workload and application spec ("instance") that could contain such information. When dealing with workloads, this information should be passed over as annotation (keptn.sh/instance or app.kubernetes.io/instance), whereby the keptn label takes precedence over the Kubernetes label, and the namespace name can be taken as a fallback. In applications, this can be set explicitly, if not set, we should fallback to the namespace in the ApplicationVersion.

[grabnerandi]

Hi. Could you give an example on the instance? Would this be used to e.g: distinguish blue/green or canary vs main?

[thschue]

For example, you might have the same application in a dev, hardening or production environment, and all of them are reporting to the same monitoring instance. The instance field should help you differentiate the metrics and traces per environment.

I didn't think of blue-green deployments when writing this issue, but this might help to distinguish between the blue and green environments. I think this is a thing we can follow up on, great idea!

[thschue]

I think we should rethink this issue when dealing with keptn/enhancement-proposals#90.

Therefore, closing this issue for the moment.