Asking for some architectural feedback

[Daniel Andrei Mincă]

Greetings,

I’ve currently deployed GitLab on 2 separate Kubernetes clusters, one’s staging the other’s production. As some may probably know, GitLab itself requires some external components as well to be provisioned such as an object store, a relational database and a cache server.

The above I’ve provisioned via Helm ( helm upgrade —install -f staging/prod-values.yaml ) from a CI/CD pipeline.

My current concern is that I’ve got to serve production traffic for at least 200 users and the current setup is not failure-proof, human-error safe and so on.

I’ve been observing kapitan.dev for at least 2 years now and today when I checked again the docus, Helm support in Kapitan is out of beta (correct me if I’m wrong).

Additionally, I also want to benefit from auto-generated docus, which kapitan would definitely support here.

How is everyone tackling this situation of let’s say, manifests validation before they reach Production, auto-generated documentation (or documentation as a general topic with Kapitan), would you be kind to give me some pointers here.

Thank you so much

Daniel-Andrei Minca

Daniel-Andrei Minca