Pods
Long-running container workloads on Atlas. Per-second billed.
What it is
Pods is the primitive behind GPU/CPU/bare-metal compute. Where functions are stateless and short-lived, pods stay around (up to 30 days) and can mount block volumes, parallel file systems and elastic IPs.
When to use it
- Anything that doesn't fit the function shape.
- Workloads that need a real network and a stable IP.
- Workloads that need a parallel file system.
Quickstart
asc pods run --cpu 8 --memory 32Gi --image my/api:latest --port 8080 --eippod = client.pods.run(image='my/api:latest', cpu=8, memory='32Gi', ports=[8080], eip=True)const pod = await client.pods.run({ image: 'my/api:latest', cpu: 8, memory: '32Gi', ports: [8080], eip: true });Limits & quotas
| Limit | Default | Burst | Notes |
|---|---|---|---|
| Per-second billed | Yes | — | |
| Max lifetime | 30 days | — | Use Agents for indefinite |
| Max CPU | 192 vCPU | — |
Pricing
See pricing. Pay-as-you-go, billed monthly via Stripe.
API surface
POST /v1/podsGET /v1/pods/{id}
Required scope(s): pods:read, pods:write. See Scopes.
Security
All access is authenticated and scoped. See Auth & scopes and Network controls.