This message was deleted.

I can only share what I have done, I think it is reasonable, but I am not an expert. I just use Fleet for CD and leave CI to the pipelining. Containers are built and stashed in a registry. Right now helm charts I'm referencing are directly pulled from git, but I'm looking at building and publishing them to the registry if/where it makes sense. I have a gitRepo that contains multiple paths for different applications, each application has a fleet.yaml that points to the helm chart in question. It also contains the values variations for my different clusters (prod, test, etc.). The gitRepo targets all clusters and I use label matching to align the values as I want. I have two gitRepos - one for the actual applications and one for the core operational things I need to curate on all my clusters (telemetry etc.) I let fleet pull/poll, and I'm not yet using the loop to watch for container updates and commit changes back to git.... The only thing I'm aware of is that updating the gitrepo for one application means changes are scanned for all of the apps in that gitrepo....so if you have many apps you may want to divide things. For this reason alone I have segmented out some dev stuff with specific cluster targeting and limited paths for the application I'm working on at that moment.

Hello Ed, thanks for the reply. I am currently trying to reuse the same helm chart for multiple deployments in the same cluster, hence same cluster with dev, staging, prod namespaces. Have you found a way to do this? It seems to me that the helm chart always has to be in a subfolder below the fleet chart in order to work. This means, I would need to have multiple copies of the helm chart. This feels weird to me. Would be great if you could explain how the repository structure and fleet.yamls would have to look like.