Skip to content
Thyme Docs

Gate usage and billing

Gate reports requests and compute units (CU) separately. Request counts describe traffic volume. Compute units are the metered usage used by the RPC plan; different methods can contribute different amounts of CU.

Current billing cycle

The Usage page shows:

  • The workspace's RPC plan and current cycle.
  • Total CU used and any included free allowance.
  • Billable CU after the allowance.
  • The configured price per million CU and estimated usage cost when a price is present.

The usage estimate is calculated as:

billable CU = max(0, used CU - free allowance CU)
estimated usage cost = billable CU × configured price per CU

The displayed amount is a usage estimate, not a final invoice total. Use the workspace's current plan and billing information for actual rates, included allowances, and any other charges. The billing cycle is derived from the subscription; it need not begin on the first day of the calendar month.

Charts and request analytics

Usage charts can show the last 24 hours, 7 days, or 30 days. The 24-hour chart uses hourly buckets; longer ranges use daily buckets. Chart labels use UTC. The method breakdown shows CU consumption by method, with an other group for remaining methods.

Changing the chart range does not change the billing-cycle total above it. A last-30-days chart and a subscription-cycle figure can cover different time windows.

The Dashboard provides request activity and method breakdowns. Endpoints shows recent per-chain traffic. API Keys shows key-level requests in the last 24 hours and last-use time. Some dashboard and billing aggregates are workspace-wide even when a particular project or key is selected elsewhere.

Limits and suspension

RPC access depends on the active service subscription and billing state. Depending on the configured plan, a quota can be a hard cap or support overage. A suspended billing state can stop requests. Creating another key in the same workspace does not create another independent workspace allowance.

No single request-rate, CU-per-method, batch-size, archive-depth, or log-range limit is guaranteed by the console alone. Check the configured plan and the response from the specific endpoint before sizing a workload around those limits.

Gate charges versus gas

Gate usage measures RPC service consumption. It does not represent the native gas paid when a transaction executes on-chain.

Flow compute, Gate RPC usage, and Lift gas sponsorship are separate concepts. The current Lift Sepolia pilot's zero-dBank gas charge does not imply that all other service usage is free.

Reduce unnecessary usage

Cache stable reads when appropriate, reuse clients, avoid polling more frequently than the application needs, and query event logs over bounded ranges. Treat a burst of retries as additional traffic: use bounded backoff and investigate persistent errors instead of retrying indefinitely.