Operator Configuration Reference

The k8up operator can be configured in two ways:

Per namespace backups. Optimal for shared clusters.
Global settings with namespaced schedules. Optimal for private clusters.

Arguments and Environment Variables

You need to define BACKUP_OPERATOR_NAMESPACE (or --operator-namespace respectively), but everything else can be left to their default values.

See k8up operator --help for all the options:

NAME:
   k8up operator

USAGE:
   k8up operator [command options] [arguments...]

DESCRIPTION:
   Start k8up in operator mode

OPTIONS:
   --annotation value                                  the annotation to be used for filtering (default: "k8up.io/backup") [$BACKUP_ANNOTATION]
   --backupcommandannotation value                     set the annotation name that identify the backup commands on Pods (default: "k8up.io/backupcommand") [$BACKUP_BACKUPCOMMANDANNOTATION]
   --fileextensionannotation value                     set the annotation name where the file extension is stored for backup commands (default: "k8up.io/file-extension") [$BACKUP_FILEEXTENSIONANNOTATION]
   --global-failed-jobs-history-limit value            set the number of old, failed jobs to keep when cleaning up, applies to all job types (default: 3) [$BACKUP_GLOBAL_FAILED_JOBS_HISTORY_LIMIT]
   --global-successful-jobs-history-limit value        set the number of old, successful jobs to keep when cleaning up, applies to all job types (default: 3) [$BACKUP_GLOBAL_SUCCESSFUL_JOBS_HISTORY_LIMIT]
   --global-concurrent-archive-jobs-limit value        set the limit of concurrent archive jobs (default: unlimited) [$BACKUP_GLOBAL_CONCURRENT_ARCHIVE_JOBS_LIMIT]
   --global-concurrent-backup-jobs-limit value         set the limit of concurrent backup jobs (default: unlimited) [$BACKUP_GLOBAL_CONCURRENT_BACKUP_JOBS_LIMIT]
   --global-concurrent-check-jobs-limit value          set the limit of concurrent check jobs (default: unlimited) [$BACKUP_GLOBAL_CONCURRENT_CHECK_JOBS_LIMIT]
   --global-concurrent-prune-jobs-limit value          set the limit of concurrent prune jobs (default: unlimited) [$BACKUP_GLOBAL_CONCURRENT_PRUNE_JOBS_LIMIT]
   --global-concurrent-restore-jobs-limit value        set the limit of concurrent restore jobs (default: unlimited) [$BACKUP_GLOBAL_CONCURRENT_RESTORE_JOBS_LIMIT]
   --globalrestores3accesskeyid value                  set the global restore S3 accessKeyID for restores [$BACKUP_GLOBALRESTORES3ACCESKEYID, $BACKUP_GLOBALRESTORES3ACCESSKEYID]
   --globalrestores3bucket value                       set the global restore S3 bucket for restores [$BACKUP_GLOBALRESTORES3BUCKET]
   --globalrestores3endpoint value                     set the global restore S3 endpoint for the restores (needs the scheme 'http' or 'https') [$BACKUP_GLOBALRESTORES3ENDPOINT]
   --globalrestores3secretaccesskey value              set the global restore S3 SecretAccessKey for restores [$BACKUP_GLOBALRESTORES3SECRETACCESSKEY]
   --globalrepopassword value                          set the restic repository password to be used globally [$BACKUP_GLOBALREPOPASSWORD]
   --globalaccesskeyid value                           set the S3 access key id to be used globally [$BACKUP_GLOBALACCESSKEYID]
   --globalsecretaccesskey value                       set the S3 secret access key to be used globally [$BACKUP_GLOBALSECRETACCESSKEY]
   --globals3bucket value                              set the S3 bucket to be used globally [$BACKUP_GLOBALS3BUCKET]
   --globals3endpoint value                            set the S3 endpoint to be used globally [$BACKUP_GLOBALS3ENDPOINT]
   --global-cpu-request value                          set the CPU request for scheduled jobs [$BACKUP_GLOBAL_CPU_REQUEST]
   --global-cpu-limit value                            set the CPU limit for scheduled jobs [$BACKUP_GLOBAL_CPU_LIMIT]
   --global-memory-request value                       set the memory request for scheduled jobs [$BACKUP_GLOBAL_MEMORY_REQUEST]
   --global-memory-limit value                         set the memory limit for scheduled jobs [$BACKUP_GLOBAL_MEMORY_LIMIT]
   --image value                                       URL of the restic image (default: "ghcr.io/k8up-io/k8up:latest") [$BACKUP_IMAGE]
   --command-restic value [ --command-restic value ]   The command that is executed for restic backups. (default: "/usr/local/bin/k8up", "restic") [$BACKUP_COMMAND_RESTIC]
   --restic-options value [ --restic-options value ]   Pass custom restic options in the form 'key=value,key2=value2'. See https://restic.readthedocs.io/en/stable/manual_rest.html?highlight=--option#usage-help [$BACKUP_RESTIC_OPTIONS]
   --datapath value, --mountpath value                 to which path the PVCs should get mounted in the backup container (default: "/data") [$BACKUP_DATAPATH]
   --globalstatsurl value                              set the URL to post metrics globally [$BACKUP_GLOBALSTATSURL]
   --metrics-bindaddress value                         set the bind address for the prometheus endpoint (default: ":8080") [$BACKUP_METRICS_BINDADDRESS]
   --promurl value                                     set the operator wide default prometheus push gateway (default: "http://127.0.0.1/") [$BACKUP_PROMURL]
   --restartpolicy value                               set the RestartPolicy for the backup jobs. According to https://kubernetes.io/docs/concepts/workloads/controllers/jobs-run-to-completion/, this should be 'OnFailure' for jobs that terminate (default: "OnFailure") [$BACKUP_RESTARTPOLICY]
   --podfilter value                                   the filter used to find the backup pods (default: "backupPod=true") [$BACKUP_PODFILTER]
   --podexecaccountname value, --serviceaccount value  set the service account name that should be used for the pod command execution (default: "pod-executor") [$BACKUP_PODEXECACCOUNTNAME]
   --podexecrolename value                             set the role name that should be used for pod command execution (default: "pod-executor") [$BACKUP_PODEXECROLENAME]
   --enable-leader-election                            enable leader election within the operator Pod (default: enabled) [$BACKUP_ENABLE_LEADER_ELECTION]
   --skip-pvcs-without-annotation                      skip selecting PVCs that don't have the BACKUP_ANNOTATION (default: disabled) [$BACKUP_SKIP_WITHOUT_ANNOTATION]
   --checkschedule value                               the default check schedule (default: "0 0 * * 0") [$BACKUP_CHECKSCHEDULE]
   --operator-namespace value                          set the namespace in which the K8up operator itself runs [$BACKUP_OPERATOR_NAMESPACE]
   --vardir value                                      the var data dir for read/write k8up data or temp file in the backup pod (default: "/k8up") [$VAR_DIR]
   --help, -h                                          show help (default: false)

Global Settings

Each variable starting with BACKUP_GLOBAL* can be used to declare a global default for all namespaces. For example, if you configure the S3 bucket and credentials here, you won’t have to specify them in the Schedule or Backup resource definitions.

It is always possible to overwrite the global settings. Simply declare the specific setting in the relevant resource definition and it will be applied instead of the global default.