From a8162cb0257d72c8d7111c2b703674a9fff0ba9a Mon Sep 17 00:00:00 2001
From: tomholford <tomholford@users.noreply.github.com>
Date: Thu, 23 Apr 2026 11:12:42 -0700
Subject: [PATCH] caddyhttp: micro-benchmark HTTP/3 server construction
 with/without WebTransport
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Adds a hermetic pair of benchmarks on buildHTTP3Server to provide
quantitative evidence for the claim that deployments with
enable_webtransport=false pay no cost for the WebTransport feature.

Results on Apple M4, go1.25, -count=5:

  BenchmarkBuildHTTP3Server_WebTransportOff   ~70 ns/op   392 B/op   3 allocs/op
  BenchmarkBuildHTTP3Server_WebTransportOn   ~144 ns/op   600 B/op   6 allocs/op

The Off path is about half the cost on every dimension, confirming
that the work skipped when the flag is false is the
webtransport.ConfigureHTTP3Server call plus EnableStreamResetPartialDelivery.
Absolute cost is a one-time per-server setup so either branch is
negligible in practice, but the asymmetry locks in a regression guard:
a future refactor that accidentally re-enables the WT configuration
unconditionally would show up as a jump in the Off numbers.

This benchmark does not exercise the per-stream dispatch cost inside
webtransport.Server.ServeQUICConn — that would require a full QUIC
setup to measure in isolation and is follow-up work.
---
 modules/caddyhttp/server_bench_test.go | 55 ++++++++++++++++++++++++++
 1 file changed, 55 insertions(+)
 create mode 100644 modules/caddyhttp/server_bench_test.go

diff --git a/modules/caddyhttp/server_bench_test.go b/modules/caddyhttp/server_bench_test.go
new file mode 100644
index 000000000..fc1d5248d
--- /dev/null
+++ b/modules/caddyhttp/server_bench_test.go
@@ -0,0 +1,55 @@
+// Copyright 2015 Matthew Holt and The Caddy Authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+package caddyhttp
+
+import (
+	"crypto/tls"
+	"testing"
+)
+
+// BenchmarkBuildHTTP3Server_WebTransportOff measures the cost of
+// constructing the HTTP/3 server for a deployment that does NOT opt in
+// to WebTransport. This is the primary cost comparison steadytao asked
+// about: any non-zero delta vs. pre-WebTransport Caddy would need to be
+// justified, and the implementation is structured so the delta is
+// exactly the branch check on EnableWebTransport.
+//
+// This benchmark does not exercise the per-stream dispatch cost
+// (which is inside webtransport-go / quic-go and would require a full
+// QUIC setup to measure in isolation). The meaningful regression guard
+// is whether buildHTTP3Server with the flag off does the same work
+// as on pre-PR master.
+func BenchmarkBuildHTTP3Server_WebTransportOff(b *testing.B) {
+	s := &Server{}
+	tlsCfg := &tls.Config{}
+	b.ResetTimer()
+	for b.Loop() {
+		_ = s.buildHTTP3Server(tlsCfg)
+	}
+}
+
+// BenchmarkBuildHTTP3Server_WebTransportOn measures the same
+// construction with WebTransport enabled. The cost difference vs. the
+// Off variant is the one-time setup webtransport.ConfigureHTTP3Server
+// performs (AdditionalSettings, ConnContext, EnableDatagrams, etc.)
+// plus setting EnableStreamResetPartialDelivery on the QUIC config.
+func BenchmarkBuildHTTP3Server_WebTransportOn(b *testing.B) {
+	s := &Server{EnableWebTransport: true}
+	tlsCfg := &tls.Config{}
+	b.ResetTimer()
+	for b.Loop() {
+		_ = s.buildHTTP3Server(tlsCfg)
+	}
+}