rp serverless

Serverless endpoints: deploy, scale, inspect and invoke. An endpoint runs a container image on demand behind a queue or a load balancer, scaling a pool of workers between a minimum and a maximum. The image comes from a template or a Hub listing rather than from flags, and jobs are submitted on the data plane with rp serverless run.

rp serverless <verb> [flags]

Commands

• rp serverless create — Create a serverless endpoint from a template or a Hub listing.
• rp serverless list — List your endpoints: id, name, worker bounds and idle timeout.
• rp serverless get — Show one endpoint's full record and scaling config.
• rp serverless update — Change an endpoint's workers, GPU pool, registry, template, name, or scaling.
• rp serverless delete — Delete a serverless endpoint permanently.
• rp serverless scale — Set an endpoint's worker bounds and idle timeout in one call.
• rp serverless workers — Show an endpoint's live workers: ids, states, placement, versions.
• rp serverless releases — Show an endpoint's release history and rollout.
• rp serverless logs — Stream one worker's container and system logs live.
• rp serverless run — Submit a job to a deployed endpoint on the data plane.
• rp serverless status — Check the status of a job submitted earlier on a deployed endpoint.
• rp serverless health — Show a deployed endpoint's operational health: worker and job counts.