Resource Limits for Teams and Products
Last verified 30 Sep 2026
To manage compute capacity and risks, and ensure fair resource distribution, DigitalOcean sets resource limits. Each resource limit caps a product, such as the number of Droplets, GPU Droplets, inference usage, database clusters, and so on. Limits apply per team or per product.
As a team owner or resource modifier, you can view resource usage and increase limits for products in your team.
Organization owners and admins can also view and adjust limits for every team in their organization.
See Droplet plans for details on Droplet choices and plans we offer.
| Tier | Droplet Count | Droplet Types | GPU Droplets |
|---|---|---|---|
| Tier 1 | 3 | Basic Shared CPU Droplets, up to $48 | 0 |
| Tier 2 | 10 | Basic Shared CPU Droplets, up to $56 Some Dedicated CPU Droplets, up to $84 |
1 |
| Tier 3 | 15 | All Basic Shared CPU Droplets Larger selection of Dedicated CPU Droplets, up to $163 |
2 |
| Tier 4 | 25 | All Droplets | 4 |
| Tier 5 | 100 | All Droplets | 16 |
- Quotas are scoped per account.
-
Serverless inference is limited by account tier and has the following limits:
Tier Requests Per Minute (RPM) Tokens Per Minute (TPM) Playground (tokens per day) Image Generation (requests per day) Video Generation (requests per day) Tier 1 Up to 120 Up to 1.5M 50K 500 80 Tier 2 Up to 120 Up to 3M 50K 1,000 80 Tier 3 Up to 180 Up to 5M 100K 2,500 400 Tier 4 Up to 360 Up to 12M 250K 10,000 750 Tier 5 Up to 1,200 Up to 25M 500K 25,000 1,000 Enterprise Contact support Contact support Contact support Contact support Contact support Rate limits vary by model. The RPM and TPM values are the maximums available at each tier, and some models carry lower limits. Token limits (TPM) apply to each model individually, and request limits (RPM) are shared across groups of similar models. The exact limit values for any model are returned in the response headers of every request and include rate-limit headers with your current limits for the model, how much you have left, and when each limit resets. These headers are provided to help you manage usage and may briefly lag actual state. Your limits never exceed the maximum published values in the table above. If a specific workload needs more than your tier allows, contact support to request an increase for a specific model.
These limits are maximums on shared capacity, not guarantees. DigitalOcean provides serverless inference on a commercially reasonable effort basis. Operating within your published limits does not guarantee that any given request will be served, or any particular throughput, latency, or availability. When capacity is constrained (including capacity for a specific model, model family, or third-party provider), requests may be queued, throttled, or rejected (for example, with a 503 or “overloaded” response) even when you are within your limits, and this is not a service defect. DigitalOcean may change these limits, tiers, and model classifications at any time; where a change would materially reduce a paying customer’s limits, we aim to provide advance notice.
All tiers, except Tier 1 and Tier 2, have access to all models. Tier 1 and Tier 2 do not have access to any Anthropic models and OpenAI models (except gpt-oss-120b and gpt-oss-20b).
See Inference Limits for additional details.
Regional Load Balancer
| Tier | Count | Nodes Per Instance |
|---|---|---|
| Tier 1 | 2 | 3 |
| Tier 2 | 5 | 6 |
| Tier 3 | 15 | 25 |
| Tier 4 | 15 | 25 |
| Tier 5 | 50 | 60 |
Network Regional Load Balancer
| Tier | Nodes Per Instance | Autoscaler |
|---|---|---|
| Tier 1 | 3 | 1 |
| Tier 2 | 6 | 2 |
| Tier 3 | 10 | 3 |
| Tier 4 | 25 | 5 |
| Tier 5 | 100 | 10 |
Global Load Balancer(GLB)
| Tier | Count | Certificates |
|---|---|---|
| Tier 1 | 0 | 10 |
| Tier 2 | 1 | 20 |
| Tier 3 | 1 | 50 |
| Tier 4 | 2 | 200 |
| Tier 5 | 50 | 200 |
| Product | Count | |
|---|---|---|
| Domains | Tier 1: 6 Tier 2: 15 Tier 3: 30 Tier 4: 100 Tier 5: 500 |
|
| VPC Peering | Tier 1: 0 Tier 2: 1 Tier 3: 1 Tier 4: 5 Tier 5: 10 |
|
| Reserved (Floating) IP | Tier 1: 3 Tier 2: 10 Tier 3: 15 Tier 4: 25 Tier 5: 75 |
|
| Reserved IPv6 | Tier 1: 6 Tier 2: 15 Tier 3: 30 Tier 4: 100 Tier 5: 500 |
|
| BYOIP | Tier 1: 0 Tier 2: 0 Tier 3: 0 Tier 4: 0 Tier 5: 2 |
|
| VPC Partner Network Connect | Tier 1: 2 Tier 2: 4 Tier 3: 6 Tier 4: 10 Tier 5: 10 Bandwidth (Mbps): Tier 1: 50,000 Tier 2: 50,000 Tier 3: 50,000 Tier 4: 50,000 Tier 5: 50,000 |
| Tier | Global Capacity TiB | Size per Volume TiB | Count |
|---|---|---|---|
| Tier 1 | 1 | ~0.1 (100 GiB) | 5,000 |
| Tier 2 | 2 | ~0.1 (100 GiB) | 5,000 |
| Tier 3 | 3 | 3 | 5,000 |
| Tier 4 | 16 | 16 | 5,000 |
| Tier 5 | 25 | 16 | 5,000 |