Create a Job
Step-by-step guide to define, launch, and monitor batch jobs on Outpost.
Prerequisites
- An Outpost account
- The Outpost CLI installed (
brew install outpostkit) - A training script or data processing pipeline ready to run
Launch a job from the CLI
The simplest way to run a job:
Available flags:
| Flag | Description |
|---|---|
--name / -n | Job name (required) |
--cloud | Cloud provider (e.g. aws, azure) |
--region | Cloud region (e.g. us-east-1) |
--gpus | GPU type and count, e.g. A10G or A100:4 |
--cpus | Number of vCPUs |
--memory | Memory in GB |
--disk-size | Disk size in GB |
--command | Command to execute (required) |
--image | Container image |
--spot | Use spot instances |
--config / -c | Config file |
-s, --namespace | Namespace (global flag) |
Outpost provisions the instance, runs your command, streams logs, and terminates the instance when the command exits.
Monitor a running job
Job statuses:
| Status | Description |
|---|---|
PENDING | Instance is being provisioned |
SETTING_UP | Setup phase is running |
RUNNING | Main command is executing |
SUCCEEDED | Completed with exit code 0 |
FAILED | Exited with non-zero code |
CANCELLED | Manually cancelled |
Launch from the dashboard
- Navigate to + New > Job from the dashboard.
- Select GPU type, cloud provider, and region.
- Enter your run command and any environment variables.
- Click Create Job.
Logs stream in real time on the job detail page.
Cancel or delete a job
Tip
Save checkpoints periodically during training. If using spot instances and the instance is preempted, your job can resume from the latest checkpoint.
Next steps
- Jobs Overview — spot instances, multi-node distribution, cost optimization
- Services — deploy long-running inference endpoints
- CLI Reference — Compute Commands — full command documentation
Previous → Overview
Next Overview →