Score:0

Best practice for gitops with helm

cn flag

I want to follow “gitops” for deployment — the state given environment should be in is checked in version control in appropriate branch and then the CD server applies the configuration. With plain kubernetes manifests the deployment definition just selects the appropriate kubernetes connection and calls kubectl apply and everything else is in the files (or the name).

But with helm the values are obviously stored in a file, but besides that there is the repository, chart and version to keep track of. I can accompany each values .yaml file with a shell script issuing the appropriate helm repo add+helm repo update+helm upgrade --install, but it feels ad-hoc an unobvious.

Is there some standard(ish) way to note down the parameters for full helm deployment? The intent is to keep the CD job definition simple and the whole thing easy to understand for the next guy who has to change something.

I've been considering the helm-controller, but it has two downsides

  • There is an open issue #100 for using authenticated repositories.
  • I am not sure how to wait for the installation to complete to get correct status from the deployment job.

so for the time being I'd prefer running helm on the CD server.

mangohost

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.