AI agents: use the documentation index at llms.txt to locate machine-readable pages. This section is indexed by https://docs.particular.net/nservicebus/llms.txt. The markdown version of this page is served as plain text. An MCP server at /mcp serves the same content via the search_docs and read_doc tools; it is read-only and needs no credentials. Markdown versions of documentation pages are available by appending .md to the page URL. Directory URLs use index.md. They are served as text/plain because some retrieval backends reject text/markdown.

Scale out

Due to specifics of the protocol used, the gateway is designed to run on at most one instance of each endpoint. Depending on the transport there are different strategies for designating a gateway hosting endpoint instance.

Brokered transports

With brokered transports (RabbitMQ, SQL Server, Azure, and MSMQ scaled-out with the unified scalability model) or bus transports with sender-side distribution, all instances are equal. All can host the gateway, but one instance must be explicitly selected as the receiver of the incoming HTTP gateway traffic. Any endpoints sending to the gateway of that endpoint must be configured to use the HTTP address of the selected endpoint instance.

Use an HTTP load balancer to avoid hard-coding individual endpoint instance gateway HTTP addresses in the gateway senders. Configure sending endpoints to send to the load balancer and let the load balancer forward the traffic to the HTTP address of the selected endpoint instance.

Handle high availability requirements by setting the load balancer to fail over to another endpoint instance if it detects endpoint failure.

Gateway scale-out

Caveats

Callbacks through the gateway are not supported on scaled-out endpoints.

List of Samples

In this article