Has anyone used fleet to deploy the Rancher applications (https://git.rancher.io/charts) across multiple downstream clusters? I'm particularly interested in trying to roll out Rancher CIS Benchmarking.

Hi, We already use Fleet to install all the Rancher charts. Here is an example :

defaultNamespace: cattle-gatekeeper-system
helm:
  repo: <https://charts.rancher.io>
  chart: rancher-gatekeeper-crd
  version: 101.0.0+up3.9.0
  releaseName: rancher-gatekeeper-crd

defaultNamespace: cattle-gatekeeper-system
helm:
  repo: <https://charts.rancher.io>
  chart: rancher-gatekeeper
  version: 101.0.0+up3.9.0
  releaseName: rancher-gatekeeper
  values:
    experimentalEnableMutation: true
    auditMatchKindOnly: true
    auditFromCache: true
dependsOn:
  - name: opa-gatekeeper-rancher-gatekeeper-crd

What would be the best approach to customize an application across different clusters? Let's say I want to deploy a CSI driver but every cluster needs to use their own backend storage path. Creating a Kustomize setup for every cluster seems like a bad idea.

You can use a targetcustomization, where you can pass in customised values for every cluster https://fleet.rancher.io/gitrepo-targets

That worked well with one application (chart) and many clusters, thanks! But what about having multiple charts over many clusters? I don't want to repeat the targetcustomizations again for each chart, is there a way to do it only once? (Something like this example, but with charts instead of yaml manifests) In the customizations, I'm changing the helm chart values only

Fleet itself is based on helm. Helm can have dependent charts. If you can create a helm chart that follows the dependent charts model as well as the values then yes. You can do what you want

Yep, I got it, thanks.