manage-runners runner autoscaling set
Create or change a runner's scaling policy, which adds paid Hetzner or DigitalOcean burst runners while the runner is busy; unchanged settings and omitted tokens are kept
| Required scopes | runners:read, runners:write |
|---|---|
| Required flags | none |
| Changes state | yes |
| Destructive | no |
| Paid resource effect | create |
| Confirmation | paid_resource |
| Non-interactive use | supported |
This command may create paid Hetzner or DigitalOcean resources. Read confirmations before running it non-interactively.
Usage
Create or change a runner's scaling policy, which adds paid Hetzner or DigitalOcean burst runners while the runner is busy; unchanged settings and omitted tokens are kept
Usage:
manage-runners runner autoscaling set [flags]
Flags:
--cache string S3 cache id or name for burst runners (external GitLab anchors); a name needs runners:read
--concurrency int burst runner concurrency, 1 to 100, for GitHub runners 1 to 32 (external anchors)
--digitalocean-connection string DigitalOcean connection id for burst runners on DigitalOcean instead of --provider-token-env (external anchors, and runners whose burst runners use DigitalOcean as another provider); see provider digitalocean connections
--enabled[=true|false] whether the policy scales; a new policy is enabled unless --enabled=false; an existing one keeps its state (default true)
--executor string burst runner executor: docker, dind or shell (external GitLab anchors)
--github-owner string GitHub user or organization the burst runners register with, the anchor's (external GitHub anchors); with --github-repository and --github-runner-group it replaces the stored scope
--github-repository string GitHub repository of the owner (external GitHub anchors); omit for an organization runner
--github-runner-group string runner group of an organization runner (external GitHub anchors); omit for the default group
--github-token-env string environment variable containing a fine-grained GitHub token for the burst runners (external GitHub anchors): Administration read and write for a repository, Self-hosted runners read and write for an organization; omit it to keep the stored token
--gitlab-host string GitLab host the burst runners register with (external GitLab anchors)
--hetzner-token string saved Hetzner token id for burst runners on Hetzner instead of --hetzner-token-env (external anchors, and runners whose burst runners use Hetzner as another provider); see provider hetzner tokens
--hetzner-token-env string environment variable containing the Hetzner token for burst runners (external anchors, and runners whose burst runners use Hetzner as another provider); omit it to keep the stored token
--id string anchor id: a runner or an external runner
--label stringArray burst runner label (external anchors), for GitHub runners also its GitHub label; repeat for several; the given labels replace all, and --label '' removes them
--location string burst runner location of the cloud provider (external anchors, and runners whose burst runners use another provider)
--max int maximum burst runners, 1 to 20; required for a new policy
--name string exact runner or external runner name
--no-cache[=true|false] burst runners use no cache (external GitLab anchors)
--no-send-to[=true|false] burst runners send to no Metrics connection (external anchors)
--preset string scale on free job slots like the dashboard's presets: eager (add when at most 1 job slot is free for 2 min; for anchors with 2 or more job slots), balanced (none free for 3 min) or conservative (none free for 5 min); replaces the rules
--product-id string burst runner product (external anchors, and runners whose burst runners use another provider)
--profile string configuration profile
--provider string cloud provider of the burst runners: hetzner|digitalocean; omit it to keep the stored provider, for a new policy the runner's own, or hetzner for an external runner; on the runner's own provider they copy the runner, another provider needs --product-id, --location and its token
--provider-token-env string environment variable containing the cloud provider's token for burst runners (external anchors, and runners whose burst runners use another provider); omit it to keep the stored token; a DigitalOcean token needs the custom scopes account:read, droplet:create, droplet:read, droplet:update, droplet:delete, image:read, reserved_ip:create, reserved_ip:read, reserved_ip:update, reserved_ip:delete, ssh_key:read and tag:create
--rule stringArray custom rule, one per metric, up to 3; replaces the rules. free:K:MINUTES adds burst runners when at most K (0-20) job slots are free for MINUTES (2-60) and removes one when more than K would stay free without it or when the group ran no job for the scale-in window; cpu:OUT:MINUTES:IN or memory:OUT:MINUTES:IN adds them when usage stays at or above OUT% (1-100) for MINUTES and removes them when it would stay below IN% (lower than OUT); for example free:0:3 or cpu:80:5:30
--runner-token-env string environment variable containing the anchor's GitLab runner authentication token (external GitLab anchors); omit it to keep the stored token
--scale-in-window int minutes the load must stay low before a burst runner is removed, 1 to 120 (new policy: 10)
--send-to string Metrics connection id or name burst runners send to (external anchors); a name needs runners:read
--ssh-key stringArray SSH key id or name of the cloud provider for burst runners (external anchors, and runners whose burst runners use another provider); a name is looked up with --hetzner-token-env, --provider-token-env, --digitalocean-connection or --hetzner-token, which needs providers:ssh-keys:read, or else with the stored token; repeat for several; the given keys replace all, and --ssh-key '' removes them
--stabilization int minutes to wait after a burst runner became ready before adding more, 0 to 60 (new policy: 3)
--step int burst runners added per step, 1 to 5 (new policy: 1)
--yes[=true|false] confirm that the policy may create paid burst runners
Every command also accepts the global flags. For machine-readable
help, run manage-runners help runner autoscaling set --output json.