Rate Limits & Usage
Delivery API consumption is metered per workspace. Each plan sets a monthly request allowance and storage cap; you can track both from the dashboard.
What is metered
One Delivery API request authenticated with a wrk_… key counts as one unit against your workspace’s monthly request allowance. The count happens once per HTTP request, on success, and is scoped to the workspace — not per project or per key.
Storage
Storage is the live sum of media bytes across all projects in the workspace. Asset bandwidth is a plan dimension but is not yet measured — media is served as object keys, so real egress lives on the CDN.
When you exceed the limit
Overages are soft and fail-open by default. When usage enforcement is enabled and your request count reaches the monthly limit, the Delivery API returns 429 RATE_LIMITED; otherwise metering never blocks delivery.
Monitoring usage
The Usage page in the dashboard shows the current billing period (calendar month, UTC boundaries), requests used vs. limit, and storage used vs. limit. The counter is batched off the hot path and flushed roughly every 15 seconds, so the number you see may lag real-time by a short window.
To stay well under your allowance, cache aggressively — the Delivery API ships cache headers and tags for exactly this (see Caching), and any CDN or framework cache in front of it absorbs repeat reads.
Caching