version numbers / image tags in kedge YAML

[jstrachan]

whats the current approach for replacing version numbers in a kedge YAML? Am thinking of CI / CD pipelines that tend to generate new unique release numbers for each new version and tend to put those versions in the YAML as a version label and then use the docker image tag of the same name as the version

should we use environment variable expressions or something? e.g.

name: foo
labels:
  version: "${VERSION}"
image: "foo/bar:${VERSION}"

then we can specify it via something like...

export VERSION="1.2.3"
kedge create -f foo.yaml (or whatever the command is ;)

[surajssd]

whats the current approach for replacing version numbers in a kedge YAML?

None, as of now.

Supporting substitution of env var like that SCTM!

[jstrachan]

cool - we'd just need to be a little careful that whatever expressions we use for env var substitution we don't clash with possible OpenShift Template / Helm parameter expressions

[jstrachan]

@kadel it'd be nice to move this to 0.2.0 to raise its importance so we can try versioned YAML generation in Jenkins pipelines soon. Even if its just simple env var substitution during YAML generation?

[kadel]

0.2.0 will be released today.
We will work on this in coming sprint and it will be added in 0.3.0 release

[pradeepto]

Ingress bits can be parameterised as well as per #41

[kadel]

I was thinking about best syntax that we can use for parameters so it doesn't conflict with Helm charts variables or OpenShift templates.

We probably shouldn't use the Bash way of specifying variables (${VARIABLE}" ). This might conflict with OpenShift templates. If we will add Helm support in the future we shouldn't use Go templates syntax for variables ({{ variable }}`), as this will conflict with Helm templating.

So I was thinking that we could use [[ variable ]] But it looks a bit strange but as far as I know it shouldn't conflict with anything in OpenShift templates, Helm, Kubernetes or Bash.

[surajssd]

@kadel

[[ variable ]]

SGTM!