MaxConnectionsPerServer

MaxConnectionsPerServer is a setting in the .NET HttpClient stack that controls how many outbound TCP connections to a single server may be opened concurrently by a single HttpClient instance. The limit is applied per destination (host and port) and is enforced by the underlying handler, such as SocketsHttpHandler.

The property is configured on the HttpClientHandler or SocketsHttpHandler used by HttpClient. A typical usage is to create a handler with a specified limit and then create the client with that handler. For example: var handler = new SocketsHttpHandler { MaxConnectionsPerServer = 20 }; var client = new HttpClient(handler);

Default values for MaxConnectionsPerServer are implementation-dependent and can vary by framework and version. In practice, the default may be suitable for general workloads, but you may need to adjust it based on load testing, server capacity, and application behavior. It is important to note that the setting applies to a single HttpClient instance; reusing HttpClient across requests ensures the limit is honored consistently.

Impact and considerations: increasing the limit can improve throughput for highly concurrent workloads by allowing more simultaneous connections to a server, but it also increases resource usage (sockets, memory, and potential contention). Conversely, lowering the limit can reduce resource pressure but may throttle performance. When using HTTP/2, multiplexing reduces the need for many parallel connections, so the practical impact of changing this limit is smaller for HTTP/2—but the setting still governs how many connections may be established in non‑HTTP/2 scenarios.

Best practices include benchmarking under realistic load, tuning the value per service and environment, and reusing HttpClient instances to benefit from connection pooling. See also related concepts such as per-server connection pooling and framework-specific connection defaults.