Skip to main content
RouteMesh pools provider capacity. Capacity depends on the chain and method you call, the eligible providers, and their current limits and availability. There is no single requests-per-second figure in this guide that guarantees capacity for every route. A healthy chain-head probe does not measure sustained throughput.

Plan a workload

For a capacity assessment, share the chains and methods, routing tier, expected sustained and peak requests per second, and how long spikes last. Include relevant response sizes or batch sizes. Support can use that scope to investigate; submitting the request does not itself confirm capacity or create a throughput commitment. If you need a firm throughput commitment, ask for a human assessment of those specific routes. An unpublished limit is unknown, not unlimited. A configured limit, an observed test result, and a contractual commitment are different kinds of evidence.

What happens during bursts

When an upstream node returns HTTP 429 or 503, RouteMesh puts the pathway on cooldown. Cooldowns are per node and method. Other eligible pathways can still serve requests. If every eligible pathway is cooling down, the router returns HTTP 429 with JSON-RPC code -32005 and error.data.reason = capacity_cooldown. See Request behavior.

Prices and query limits

This routing behavior does not establish a fixed burst price or a price guarantee. Use current pricing for published route prices; ask support about a commercial commitment for your traffic mix. The public pricing endpoint’s own per-IP request limit is separate from RPC capacity. For eth_getLogs, the block-span ceiling and the returned route-specific hint are query limits, not requests-per-second limits. See eth_getLogs limits.