Hi all, Anyone has some idea how to resolve the f...
# rke2
k
Hi all, Anyone has some idea how to resolve the following challange; We deploy RKE2 clusters using Rancher using some user specific configuration in the rke2-charts (e.g. rke2-cilium), like enabling metrics with ServiceMonitors. But in our case the ServiceMonitor CRD is still missing when RKE2 installs Cilium, and thus this installation fails. I've tried to use a HelmChart via the Additional Manifests, but it appears that these are only run after the initial bootstrap. Any idea how I can install the ServiceMonitor CRD before the helm-install-cilium Job is started?
c
How exactly are you injecting the servicemonitor definitions into the Cilium helm chart?
k
I’ve configured the rke2-cilium helm chart by enabling the value prometheus.serviceMonitor.enabled. So I’m basically looking for a way to install the ServiceMonitor CRD. I’ve tried to use the additional manifests installing the HelmChart from Prometheus-operator-crds. But these are only applied after the rke2-cilium chart, but that one fails because the ServiceMonitor CRD is missing.
c
You could set it as a bootstrap chart so that it gets installed earlier
That would probably be my recommendation
k
How can I configure such a bootstrap chart, as that sound like an option.
Because I only see the options to set chart values (the rke2-* charts) or additional chart (which can be any resource).
c
Set
bootstrap: true
in the helmchart spec. See the field documentation here. https://docs.rke2.io/helm#using-the-helm-crd
k
Ah cool. Thanks. Will test this tomorrow. Thank you for your support!
(And I should better read the docs. But overlooked the bootstrap value)