A reference list of, mostly internal, configuration options.
The Helm charts accept, at least, the options as shown with their default in
The controllers can be started with these environment variables:
CATTLE_DEV_MODE- used to debug wrangler, not usable
FLEET_CLUSTER_ENQUEUE_DELAY- tune how often non-ready clusters are checked
FLEET_CPU_PPROF_PERIOD- used to turn on performance profiling
In cluster configuration for the agent and fleet manager. Changing these can lead to full re-deployments.
The config struct is used in both config maps:
Labels used by fleet:
fleet.cattle.io/agent=true- NodeSelector label for agent's deployment affinity setting
fleet.cattle.io/non-managed-agent- managed agent bundle won't target Clusters with this label
fleet.cattle.io/repo-name- used on Bundle to reference the git repo resource
fleet.cattle.io/bundle-namespace- used on BundleDeployment to reference the Bundle resource
fleet.cattle.io/bundle-name- used on BundleDeployment to reference the Bundle resource
fleet.cattle.io/managed=true- cluster namespaces with this label will be cleaned up. Other resources will be cleaned up if it is in a label. Used in Rancher to identify fleet namespaces.
Annotations used by fleet:
fleet.cattle.io/cluster-namespace- if present on a Cluster, the namespace won't be cleaned up
fleet.cattle.io/cluster-namespace- used on a cluster namespace to reference the cluster registration namespace
fleet.cattle.io/managed- appears unused