MCP rate limits
Cap each agent's traffic through the security definition it authenticates with.
An agent caught in a loop can call a tool endpoint far faster than any human client would. Rate limits cap that traffic, per agent, before any tool runs.
Limits per identity
Limits for MCP gateways are identity-aware - they are configured on the security definition an agent authenticates with, not on the gateway. The limits configured on the definition apply to everything its holder calls, MCP gateways and REST channels alike, as one shared budget.
This means:
- Two agents calling the same gateway with different credentials each have their own limit.
- One agent calling three gateways with one credential has a single limit across all three.
- An agent that also calls REST APIs with the same credential draws from the same budget there too.
Limits are configured in the dashboard on the security definition - API key, Basic Auth or bearer token - and everything about rules, per-second bursts, per-day and per-month hard caps, CIDR address lists and time ranges is described on the rate limiting and firewall page. It all applies to MCP traffic unchanged.
Rejected requests
A request over the limit is refused before it reaches any tool, in the same way the rate limiting page describes for other clients, and with the audit log on, nothing is recorded for traffic that never reached the gateway - the refusal happens a layer earlier.
Sizing tip: Agents retry aggressively when a tool call fails, so prefer burst limits that absorb short spikes plus a daily hard cap, over a single tight per-second limit - the rate limiting guide shows how to combine both on one rule.
See also
| Feature | What it does |
|---|---|
| Security | The security definitions that carry the limits |
| Rate limiting and firewall | Rules, bursts, hard caps, CIDR lists and time ranges in full |
| Cost and limits | Every limit that applies to AI traffic, in one place |
| Audit log | What is recorded for requests that reach the gateway |