Instance Sizes

An instance is one running replica of your app. Its size determines how much CPU and memory that replica receives. If your app runs multiple replicas, each replica gets the full resources shown for the selected size in the table below.

Available Sizes

Instance size	CPU per replica	Memory per replica	Beta availability
Free	0.1 CPU	500 MiB	Hobby and Pro
Micro	0.5 CPU	500 MiB	Pro beta rollout
Small	1 CPU	2 GiB	Pro beta rollout
Medium	2 CPU	4 GiB	Pro beta rollout
Large	4 CPU	8 GiB	Coming later
XLarge	8 CPU	16 GiB	Coming later
2XLarge	16 CPU	32 GiB	Coming later

Beta

During the beta, Micro, Small, and Medium are being rolled out gradually to Pro teams and are included without additional compute charges. If your Pro team doesn’t have access yet, contact us to request early access.

Large, XLarge, and 2XLarge will be released later alongside compute-based billing. Compute prices haven’t been set yet. We’ll share pricing and give advance notice before compute billing begins. See Billing for the current plan details.

Change an App’s Instance Size

1. Navigate to your app in the dashboard.
2. Select Scaling from the sidebar.
3. In Instance Size, click Change.
4. Select an instance size.
5. Click Update and redeploy.

Changing the instance size creates a new deployment with the selected CPU and memory. The current deployment continues serving traffic until the replacement is ready.

Instance sizes other than Free require a Pro team. If you select a Pro size while using Hobby, the dashboard prompts you to upgrade.

Choose an Instance Size

Start with the smallest size that comfortably handles your app’s normal workload, then use the Metrics page to watch CPU and memory usage.

• If CPU usage repeatedly reaches its limit, requests may be throttled. Move to a larger size or optimize CPU-intensive work.
• If memory usage reaches its limit, the instance can be terminated for running out of memory and restarted. Move to a larger size or investigate memory growth.

Size and replica count solve different problems. A larger instance gives each replica more CPU and memory, while autoscaling adds or removes replicas in response to traffic. The total resources available to your app depend on both settings.