fleet dump

Dump cluster data into an archive.

Overview

Dump data from a cluster into an archive, including:

  • Fleet-managed resources

  • Kubernetes events

  • Metrics exposed by Fleet controllers

fleet dump <flags>
[source,text]

Flags

  -p <path>         Path to the archive file to generate
[source,text]

Output

Contents

The dump command produces a .tgz archive containing the following data:

  • Raw Fleet resources, in YAML format:

    • Bundles

    • BundleDeployments

    • BundleNamespaceMappings

    • Clusters

    • ClusterGroups

    • GitRepos

    • GitRepoRestrictions

    • HelmOps

  • Kubernetes events for the following namespaces:

    • cattle-fleet-system

    • cattle-fleet-local-system

    • default

    • kube-system

    • each namespace containing a Fleet Cluster resource

  • Metrics exposed by monitoring-* services living in the cattle-fleet-system namespace. This typically includes monitoring-gitjob and monitoring-fleet-controller services, as well as their sharded counterparts, if any.

Format

Raw Fleet resources are each stored in a different file, named as <kind>_<namespace>_<name>, where <kind> is one of the raw resource names mentioned in the previous section, in lowercase.

Kubernetes events are split into files by namespace, each file named as events_<namespace>. They contain one event per line, each event being encoded in JSON.

Metrics are stored into files by service, each file being named metrics_<service_name>. Their contents are GET response bodies received from each service’s /metrics endpoint, verbatim.

SEE ALSO