diff --git a/go.mod b/go.mod
index 3df52a4954f..4c01fef4635 100644
--- a/go.mod
+++ b/go.mod
@@ -30,7 +30,7 @@ require (
 	go.opentelemetry.io/otel/sdk/metric v1.19.0
 	go.uber.org/automaxprocs v1.5.3
 	go.uber.org/zap v1.26.0
-	golang.org/x/net v0.22.0
+	golang.org/x/net v0.23.0
 	golang.org/x/oauth2 v0.17.0
 	golang.org/x/sync v0.6.0
 	golang.org/x/time v0.5.0
@@ -38,14 +38,14 @@ require (
 	google.golang.org/protobuf v1.33.0
 	gopkg.in/yaml.v2 v2.4.0
 	gopkg.in/yaml.v3 v3.0.1
-	k8s.io/api v0.29.3
-	k8s.io/apiextensions-apiserver v0.29.3
-	k8s.io/apimachinery v0.29.3
-	k8s.io/client-go v0.29.3
+	k8s.io/api v0.29.4
+	k8s.io/apiextensions-apiserver v0.29.4
+	k8s.io/apimachinery v0.29.4
+	k8s.io/client-go v0.29.4
 	k8s.io/klog/v2 v2.110.1
 	k8s.io/utils v0.0.0-20230726121419-3b25d923346b
 	oras.land/oras-go v1.2.5
-	sigs.k8s.io/controller-runtime v0.17.2
+	sigs.k8s.io/controller-runtime v0.17.5
 	sigs.k8s.io/release-utils v0.7.7
 	sigs.k8s.io/yaml v1.4.0
 )
@@ -148,8 +148,8 @@ require (
 	google.golang.org/genproto/googleapis/api v0.0.0-20240311132316-a219d84964c2 // indirect
 	google.golang.org/genproto/googleapis/rpc v0.0.0-20240311132316-a219d84964c2 // indirect
 	gopkg.in/inf.v0 v0.9.1 // indirect
-	k8s.io/apiserver v0.29.3 // indirect
-	k8s.io/component-base v0.29.3 // indirect
+	k8s.io/apiserver v0.29.4 // indirect
+	k8s.io/component-base v0.29.4 // indirect
 	k8s.io/kube-openapi v0.0.0-20231010175941-2dd684a91f00 // indirect
 	sigs.k8s.io/apiserver-network-proxy/konnectivity-client v0.28.0 // indirect
 	sigs.k8s.io/json v0.0.0-20221116044647-bc3834ca7abd // indirect
diff --git a/go.sum b/go.sum
index ca6693945bb..9338749810f 100644
--- a/go.sum
+++ b/go.sum
@@ -447,8 +447,8 @@ golang.org/x/net v0.0.0-20201021035429-f5854403a974/go.mod h1:sp8m0HH+o8qH0wwXwY
 golang.org/x/net v0.0.0-20201110031124-69a78807bb2b/go.mod h1:sp8m0HH+o8qH0wwXwYZr8TS3Oi6o0r6Gce1SSxlDquU=
 golang.org/x/net v0.0.0-20210226172049-e18ecbb05110/go.mod h1:m0MpNAwzfU5UDzcl9v0D8zg8gWTRqZa9RBIspLL5mdg=
 golang.org/x/net v0.0.0-20220722155237-a158d28d115b/go.mod h1:XRhObCWvk6IyKnWLug+ECip1KBveYUHfp+8e9klMJ9c=
-golang.org/x/net v0.22.0 h1:9sGLhx7iRIHEiX0oAJ3MRZMUCElJgy7Br1nO+AMN3Tc=
-golang.org/x/net v0.22.0/go.mod h1:JKghWKKOSdJwpW2GEx0Ja7fmaKnMsbu+MWVZTokSYmg=
+golang.org/x/net v0.23.0 h1:7EYJ93RZ9vYSZAIb2x3lnuvqO5zneoD6IvWjuhfxjTs=
+golang.org/x/net v0.23.0/go.mod h1:JKghWKKOSdJwpW2GEx0Ja7fmaKnMsbu+MWVZTokSYmg=
 golang.org/x/oauth2 v0.0.0-20180821212333-d2e6202438be/go.mod h1:N/0e6XlmueqKjAGxoOufVs8QHGRruUQn6yWY3a++T0U=
 golang.org/x/oauth2 v0.17.0 h1:6m3ZPmLEFdVxKKWnKq4VqZ60gutO35zm+zrAHVmHyDQ=
 golang.org/x/oauth2 v0.17.0/go.mod h1:OzPDGQiuQMguemayvdylqddI7qcD9lnSDb+1FiwQ5HA=
@@ -558,18 +558,18 @@ gotest.tools/v3 v3.4.0 h1:ZazjZUfuVeZGLAmlKKuyv3IKP5orXcwtOwDQH6YVr6o=
 gotest.tools/v3 v3.4.0/go.mod h1:CtbdzLSsqVhDgMtKsx03ird5YTGB3ar27v0u/yKBW5g=
 honnef.co/go/tools v0.0.0-20190102054323-c2f93a96b099/go.mod h1:rf3lG4BRIbNafJWhAfAdb/ePZxsR/4RtNHQocxwk9r4=
 honnef.co/go/tools v0.0.0-20190523083050-ea95bdfd59fc/go.mod h1:rf3lG4BRIbNafJWhAfAdb/ePZxsR/4RtNHQocxwk9r4=
-k8s.io/api v0.29.3 h1:2ORfZ7+bGC3YJqGpV0KSDDEVf8hdGQ6A03/50vj8pmw=
-k8s.io/api v0.29.3/go.mod h1:y2yg2NTyHUUkIoTC+phinTnEa3KFM6RZ3szxt014a80=
-k8s.io/apiextensions-apiserver v0.29.3 h1:9HF+EtZaVpFjStakF4yVufnXGPRppWFEQ87qnO91YeI=
-k8s.io/apiextensions-apiserver v0.29.3/go.mod h1:po0XiY5scnpJfFizNGo6puNU6Fq6D70UJY2Cb2KwAVc=
-k8s.io/apimachinery v0.29.3 h1:2tbx+5L7RNvqJjn7RIuIKu9XTsIZ9Z5wX2G22XAa5EU=
-k8s.io/apimachinery v0.29.3/go.mod h1:hx/S4V2PNW4OMg3WizRrHutyB5la0iCUbZym+W0EQIU=
-k8s.io/apiserver v0.29.3 h1:xR7ELlJ/BZSr2n4CnD3lfA4gzFivh0wwfNfz9L0WZcE=
-k8s.io/apiserver v0.29.3/go.mod h1:hrvXlwfRulbMbBgmWRQlFru2b/JySDpmzvQwwk4GUOs=
-k8s.io/client-go v0.29.3 h1:R/zaZbEAxqComZ9FHeQwOh3Y1ZUs7FaHKZdQtIc2WZg=
-k8s.io/client-go v0.29.3/go.mod h1:tkDisCvgPfiRpxGnOORfkljmS+UrW+WtXAy2fTvXJB0=
-k8s.io/component-base v0.29.3 h1:Oq9/nddUxlnrCuuR2K/jp6aflVvc0uDvxMzAWxnGzAo=
-k8s.io/component-base v0.29.3/go.mod h1:Yuj33XXjuOk2BAaHsIGHhCKZQAgYKhqIxIjIr2UXYio=
+k8s.io/api v0.29.4 h1:WEnF/XdxuCxdG3ayHNRR8yH3cI1B/llkWBma6bq4R3w=
+k8s.io/api v0.29.4/go.mod h1:DetSv0t4FBTcEpfA84NJV3g9a7+rSzlUHk5ADAYHUv0=
+k8s.io/apiextensions-apiserver v0.29.4 h1:M7hbuHU/ckbibR7yPbe6DyNWgTFKNmZDbdZKD8q1Smk=
+k8s.io/apiextensions-apiserver v0.29.4/go.mod h1:TTDC9fB+0kHY2rogf5hgBR03KBKCwED+GHUsXGpR7SM=
+k8s.io/apimachinery v0.29.4 h1:RaFdJiDmuKs/8cm1M6Dh1Kvyh59YQFDcFuFTSmXes6Q=
+k8s.io/apimachinery v0.29.4/go.mod h1:i3FJVwhvSp/6n8Fl4K97PJEP8C+MM+aoDq4+ZJBf70Y=
+k8s.io/apiserver v0.29.4 h1:wPwGOO58GQOpRiZu59P5eRoDcB7QtV+QBglkRiXwCiM=
+k8s.io/apiserver v0.29.4/go.mod h1:VqTF9t98HVfhKZVRohCPezsdUt9u2g3bHKftxGcXoRo=
+k8s.io/client-go v0.29.4 h1:79ytIedxVfyXV8rpH3jCBW0u+un0fxHDwX5F9K8dPR8=
+k8s.io/client-go v0.29.4/go.mod h1:kC1thZQ4zQWYwldsfI088BbK6RkxK+aF5ebV8y9Q4tk=
+k8s.io/component-base v0.29.4 h1:xeKzuuHI/1tjleu5jycDAcYbhAxeGHCQBZUY2eRIkOo=
+k8s.io/component-base v0.29.4/go.mod h1:pYjt+oEZP9gtmwSikwAJgfSBikqKX2gOqRat0QjmQt0=
 k8s.io/klog/v2 v2.110.1 h1:U/Af64HJf7FcwMcXyKm2RPM22WZzyR7OSpYj5tg3cL0=
 k8s.io/klog/v2 v2.110.1/go.mod h1:YGtd1984u+GgbuZ7e08/yBuAfKLSO0+uR1Fhi6ExXjo=
 k8s.io/kube-aggregator v0.28.1 h1:rvG4llYnQKHjj6YjjoBPEJxfD1uH0DJwkrJTNKGAaCs=
@@ -582,8 +582,8 @@ oras.land/oras-go v1.2.5 h1:XpYuAwAb0DfQsunIyMfeET92emK8km3W4yEzZvUbsTo=
 oras.land/oras-go v1.2.5/go.mod h1:PuAwRShRZCsZb7g8Ar3jKKQR/2A/qN+pkYxIOd/FAoo=
 sigs.k8s.io/apiserver-network-proxy/konnectivity-client v0.28.0 h1:TgtAeesdhpm2SGwkQasmbeqDo8th5wOBA5h/AjTKA4I=
 sigs.k8s.io/apiserver-network-proxy/konnectivity-client v0.28.0/go.mod h1:VHVDI/KrK4fjnV61bE2g3sA7tiETLn8sooImelsCx3Y=
-sigs.k8s.io/controller-runtime v0.17.2 h1:FwHwD1CTUemg0pW2otk7/U5/i5m2ymzvOXdbeGOUvw0=
-sigs.k8s.io/controller-runtime v0.17.2/go.mod h1:+MngTvIQQQhfXtwfdGw/UOQ/aIaqsYywfCINOtwMO/s=
+sigs.k8s.io/controller-runtime v0.17.5 h1:1FI9Lm7NiOOmBsgTV36/s2XrEFXnO2C4sbg/Zme72Rw=
+sigs.k8s.io/controller-runtime v0.17.5/go.mod h1:N0jpP5Lo7lMTF9aL56Z/B2oWBJjey6StQM0jRbKQXtY=
 sigs.k8s.io/json v0.0.0-20221116044647-bc3834ca7abd h1:EDPBXCAspyGV4jQlpZSudPeMmr1bNJefnuqLsRAsHZo=
 sigs.k8s.io/json v0.0.0-20221116044647-bc3834ca7abd/go.mod h1:B8JuhiUyNFVKdsE8h686QcCxMaH6HrOAZj4vswFpcB0=
 sigs.k8s.io/release-utils v0.7.7 h1:JKDOvhCk6zW8ipEOkpTGDH/mW3TI+XqtPp16aaQ79FU=
diff --git a/vendor/golang.org/x/net/http2/frame.go b/vendor/golang.org/x/net/http2/frame.go
index e2b298d8593..43557ab7e97 100644
--- a/vendor/golang.org/x/net/http2/frame.go
+++ b/vendor/golang.org/x/net/http2/frame.go
@@ -1564,6 +1564,7 @@ func (fr *Framer) readMetaFrame(hf *HeadersFrame) (*MetaHeadersFrame, error) {
 		if size > remainSize {
 			hdec.SetEmitEnabled(false)
 			mh.Truncated = true
+			remainSize = 0
 			return
 		}
 		remainSize -= size
@@ -1576,6 +1577,36 @@ func (fr *Framer) readMetaFrame(hf *HeadersFrame) (*MetaHeadersFrame, error) {
 	var hc headersOrContinuation = hf
 	for {
 		frag := hc.HeaderBlockFragment()
+
+		// Avoid parsing large amounts of headers that we will then discard.
+		// If the sender exceeds the max header list size by too much,
+		// skip parsing the fragment and close the connection.
+		//
+		// "Too much" is either any CONTINUATION frame after we've already
+		// exceeded the max header list size (in which case remainSize is 0),
+		// or a frame whose encoded size is more than twice the remaining
+		// header list bytes we're willing to accept.
+		if int64(len(frag)) > int64(2*remainSize) {
+			if VerboseLogs {
+				log.Printf("http2: header list too large")
+			}
+			// It would be nice to send a RST_STREAM before sending the GOAWAY,
+			// but the structure of the server's frame writer makes this difficult.
+			return nil, ConnectionError(ErrCodeProtocol)
+		}
+
+		// Also close the connection after any CONTINUATION frame following an
+		// invalid header, since we stop tracking the size of the headers after
+		// an invalid one.
+		if invalid != nil {
+			if VerboseLogs {
+				log.Printf("http2: invalid header: %v", invalid)
+			}
+			// It would be nice to send a RST_STREAM before sending the GOAWAY,
+			// but the structure of the server's frame writer makes this difficult.
+			return nil, ConnectionError(ErrCodeProtocol)
+		}
+
 		if _, err := hdec.Write(frag); err != nil {
 			return nil, ConnectionError(ErrCodeCompression)
 		}
diff --git a/vendor/golang.org/x/net/http2/pipe.go b/vendor/golang.org/x/net/http2/pipe.go
index 684d984fd96..3b9f06b9624 100644
--- a/vendor/golang.org/x/net/http2/pipe.go
+++ b/vendor/golang.org/x/net/http2/pipe.go
@@ -77,7 +77,10 @@ func (p *pipe) Read(d []byte) (n int, err error) {
 	}
 }
 
-var errClosedPipeWrite = errors.New("write on closed buffer")
+var (
+	errClosedPipeWrite        = errors.New("write on closed buffer")
+	errUninitializedPipeWrite = errors.New("write on uninitialized buffer")
+)
 
 // Write copies bytes from p into the buffer and wakes a reader.
 // It is an error to write more data than the buffer can hold.
@@ -91,6 +94,12 @@ func (p *pipe) Write(d []byte) (n int, err error) {
 	if p.err != nil || p.breakErr != nil {
 		return 0, errClosedPipeWrite
 	}
+	// pipe.setBuffer is never invoked, leaving the buffer uninitialized.
+	// We shouldn't try to write to an uninitialized pipe,
+	// but returning an error is better than panicking.
+	if p.b == nil {
+		return 0, errUninitializedPipeWrite
+	}
 	return p.b.Write(d)
 }
 
diff --git a/vendor/golang.org/x/net/http2/server.go b/vendor/golang.org/x/net/http2/server.go
index ae94c6408d5..ce2e8b40eee 100644
--- a/vendor/golang.org/x/net/http2/server.go
+++ b/vendor/golang.org/x/net/http2/server.go
@@ -124,6 +124,7 @@ type Server struct {
 	// IdleTimeout specifies how long until idle clients should be
 	// closed with a GOAWAY frame. PING frames are not considered
 	// activity for the purposes of IdleTimeout.
+	// If zero or negative, there is no timeout.
 	IdleTimeout time.Duration
 
 	// MaxUploadBufferPerConnection is the size of the initial flow
@@ -434,7 +435,7 @@ func (s *Server) ServeConn(c net.Conn, opts *ServeConnOpts) {
 	// passes the connection off to us with the deadline already set.
 	// Write deadlines are set per stream in serverConn.newStream.
 	// Disarm the net.Conn write deadline here.
-	if sc.hs.WriteTimeout != 0 {
+	if sc.hs.WriteTimeout > 0 {
 		sc.conn.SetWriteDeadline(time.Time{})
 	}
 
@@ -924,7 +925,7 @@ func (sc *serverConn) serve() {
 	sc.setConnState(http.StateActive)
 	sc.setConnState(http.StateIdle)
 
-	if sc.srv.IdleTimeout != 0 {
+	if sc.srv.IdleTimeout > 0 {
 		sc.idleTimer = time.AfterFunc(sc.srv.IdleTimeout, sc.onIdleTimer)
 		defer sc.idleTimer.Stop()
 	}
@@ -1637,7 +1638,7 @@ func (sc *serverConn) closeStream(st *stream, err error) {
 	delete(sc.streams, st.id)
 	if len(sc.streams) == 0 {
 		sc.setConnState(http.StateIdle)
-		if sc.srv.IdleTimeout != 0 {
+		if sc.srv.IdleTimeout > 0 {
 			sc.idleTimer.Reset(sc.srv.IdleTimeout)
 		}
 		if h1ServerKeepAlivesDisabled(sc.hs) {
@@ -2017,7 +2018,7 @@ func (sc *serverConn) processHeaders(f *MetaHeadersFrame) error {
 	// similar to how the http1 server works. Here it's
 	// technically more like the http1 Server's ReadHeaderTimeout
 	// (in Go 1.8), though. That's a more sane option anyway.
-	if sc.hs.ReadTimeout != 0 {
+	if sc.hs.ReadTimeout > 0 {
 		sc.conn.SetReadDeadline(time.Time{})
 		st.readDeadline = time.AfterFunc(sc.hs.ReadTimeout, st.onReadTimeout)
 	}
@@ -2038,7 +2039,7 @@ func (sc *serverConn) upgradeRequest(req *http.Request) {
 
 	// Disable any read deadline set by the net/http package
 	// prior to the upgrade.
-	if sc.hs.ReadTimeout != 0 {
+	if sc.hs.ReadTimeout > 0 {
 		sc.conn.SetReadDeadline(time.Time{})
 	}
 
@@ -2116,7 +2117,7 @@ func (sc *serverConn) newStream(id, pusherID uint32, state streamState) *stream
 	st.flow.conn = &sc.flow // link to conn-level counter
 	st.flow.add(sc.initialStreamSendWindowSize)
 	st.inflow.init(sc.srv.initialStreamRecvWindowSize())
-	if sc.hs.WriteTimeout != 0 {
+	if sc.hs.WriteTimeout > 0 {
 		st.writeDeadline = time.AfterFunc(sc.hs.WriteTimeout, st.onWriteTimeout)
 	}
 
diff --git a/vendor/golang.org/x/net/http2/testsync.go b/vendor/golang.org/x/net/http2/testsync.go
new file mode 100644
index 00000000000..61075bd16d3
--- /dev/null
+++ b/vendor/golang.org/x/net/http2/testsync.go
@@ -0,0 +1,331 @@
+// Copyright 2024 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+package http2
+
+import (
+	"context"
+	"sync"
+	"time"
+)
+
+// testSyncHooks coordinates goroutines in tests.
+//
+// For example, a call to ClientConn.RoundTrip involves several goroutines, including:
+//   - the goroutine running RoundTrip;
+//   - the clientStream.doRequest goroutine, which writes the request; and
+//   - the clientStream.readLoop goroutine, which reads the response.
+//
+// Using testSyncHooks, a test can start a RoundTrip and identify when all these goroutines
+// are blocked waiting for some condition such as reading the Request.Body or waiting for
+// flow control to become available.
+//
+// The testSyncHooks also manage timers and synthetic time in tests.
+// This permits us to, for example, start a request and cause it to time out waiting for
+// response headers without resorting to time.Sleep calls.
+type testSyncHooks struct {
+	// active/inactive act as a mutex and condition variable.
+	//
+	//  - neither chan contains a value: testSyncHooks is locked.
+	//  - active contains a value: unlocked, and at least one goroutine is not blocked
+	//  - inactive contains a value: unlocked, and all goroutines are blocked
+	active   chan struct{}
+	inactive chan struct{}
+
+	// goroutine counts
+	total    int                     // total goroutines
+	condwait map[*sync.Cond]int      // blocked in sync.Cond.Wait
+	blocked  []*testBlockedGoroutine // otherwise blocked
+
+	// fake time
+	now    time.Time
+	timers []*fakeTimer
+
+	// Transport testing: Report various events.
+	newclientconn func(*ClientConn)
+	newstream     func(*clientStream)
+}
+
+// testBlockedGoroutine is a blocked goroutine.
+type testBlockedGoroutine struct {
+	f  func() bool   // blocked until f returns true
+	ch chan struct{} // closed when unblocked
+}
+
+func newTestSyncHooks() *testSyncHooks {
+	h := &testSyncHooks{
+		active:   make(chan struct{}, 1),
+		inactive: make(chan struct{}, 1),
+		condwait: map[*sync.Cond]int{},
+	}
+	h.inactive <- struct{}{}
+	h.now = time.Date(2000, 1, 1, 0, 0, 0, 0, time.UTC)
+	return h
+}
+
+// lock acquires the testSyncHooks mutex.
+func (h *testSyncHooks) lock() {
+	select {
+	case <-h.active:
+	case <-h.inactive:
+	}
+}
+
+// waitInactive waits for all goroutines to become inactive.
+func (h *testSyncHooks) waitInactive() {
+	for {
+		<-h.inactive
+		if !h.unlock() {
+			break
+		}
+	}
+}
+
+// unlock releases the testSyncHooks mutex.
+// It reports whether any goroutines are active.
+func (h *testSyncHooks) unlock() (active bool) {
+	// Look for a blocked goroutine which can be unblocked.
+	blocked := h.blocked[:0]
+	unblocked := false
+	for _, b := range h.blocked {
+		if !unblocked && b.f() {
+			unblocked = true
+			close(b.ch)
+		} else {
+			blocked = append(blocked, b)
+		}
+	}
+	h.blocked = blocked
+
+	// Count goroutines blocked on condition variables.
+	condwait := 0
+	for _, count := range h.condwait {
+		condwait += count
+	}
+
+	if h.total > condwait+len(blocked) {
+		h.active <- struct{}{}
+		return true
+	} else {
+		h.inactive <- struct{}{}
+		return false
+	}
+}
+
+// goRun starts a new goroutine.
+func (h *testSyncHooks) goRun(f func()) {
+	h.lock()
+	h.total++
+	h.unlock()
+	go func() {
+		defer func() {
+			h.lock()
+			h.total--
+			h.unlock()
+		}()
+		f()
+	}()
+}
+
+// blockUntil indicates that a goroutine is blocked waiting for some condition to become true.
+// It waits until f returns true before proceeding.
+//
+// Example usage:
+//
+//	h.blockUntil(func() bool {
+//		// Is the context done yet?
+//		select {
+//		case <-ctx.Done():
+//		default:
+//			return false
+//		}
+//		return true
+//	})
+//	// Wait for the context to become done.
+//	<-ctx.Done()
+//
+// The function f passed to blockUntil must be non-blocking and idempotent.
+func (h *testSyncHooks) blockUntil(f func() bool) {
+	if f() {
+		return
+	}
+	ch := make(chan struct{})
+	h.lock()
+	h.blocked = append(h.blocked, &testBlockedGoroutine{
+		f:  f,
+		ch: ch,
+	})
+	h.unlock()
+	<-ch
+}
+
+// broadcast is sync.Cond.Broadcast.
+func (h *testSyncHooks) condBroadcast(cond *sync.Cond) {
+	h.lock()
+	delete(h.condwait, cond)
+	h.unlock()
+	cond.Broadcast()
+}
+
+// broadcast is sync.Cond.Wait.
+func (h *testSyncHooks) condWait(cond *sync.Cond) {
+	h.lock()
+	h.condwait[cond]++
+	h.unlock()
+}
+
+// newTimer creates a new fake timer.
+func (h *testSyncHooks) newTimer(d time.Duration) timer {
+	h.lock()
+	defer h.unlock()
+	t := &fakeTimer{
+		hooks: h,
+		when:  h.now.Add(d),
+		c:     make(chan time.Time),
+	}
+	h.timers = append(h.timers, t)
+	return t
+}
+
+// afterFunc creates a new fake AfterFunc timer.
+func (h *testSyncHooks) afterFunc(d time.Duration, f func()) timer {
+	h.lock()
+	defer h.unlock()
+	t := &fakeTimer{
+		hooks: h,
+		when:  h.now.Add(d),
+		f:     f,
+	}
+	h.timers = append(h.timers, t)
+	return t
+}
+
+func (h *testSyncHooks) contextWithTimeout(ctx context.Context, d time.Duration) (context.Context, context.CancelFunc) {
+	ctx, cancel := context.WithCancel(ctx)
+	t := h.afterFunc(d, cancel)
+	return ctx, func() {
+		t.Stop()
+		cancel()
+	}
+}
+
+func (h *testSyncHooks) timeUntilEvent() time.Duration {
+	h.lock()
+	defer h.unlock()
+	var next time.Time
+	for _, t := range h.timers {
+		if next.IsZero() || t.when.Before(next) {
+			next = t.when
+		}
+	}
+	if d := next.Sub(h.now); d > 0 {
+		return d
+	}
+	return 0
+}
+
+// advance advances time and causes synthetic timers to fire.
+func (h *testSyncHooks) advance(d time.Duration) {
+	h.lock()
+	defer h.unlock()
+	h.now = h.now.Add(d)
+	timers := h.timers[:0]
+	for _, t := range h.timers {
+		t := t // remove after go.mod depends on go1.22
+		t.mu.Lock()
+		switch {
+		case t.when.After(h.now):
+			timers = append(timers, t)
+		case t.when.IsZero():
+			// stopped timer
+		default:
+			t.when = time.Time{}
+			if t.c != nil {
+				close(t.c)
+			}
+			if t.f != nil {
+				h.total++
+				go func() {
+					defer func() {
+						h.lock()
+						h.total--
+						h.unlock()
+					}()
+					t.f()
+				}()
+			}
+		}
+		t.mu.Unlock()
+	}
+	h.timers = timers
+}
+
+// A timer wraps a time.Timer, or a synthetic equivalent in tests.
+// Unlike time.Timer, timer is single-use: The timer channel is closed when the timer expires.
+type timer interface {
+	C() <-chan time.Time
+	Stop() bool
+	Reset(d time.Duration) bool
+}
+
+// timeTimer implements timer using real time.
+type timeTimer struct {
+	t *time.Timer
+	c chan time.Time
+}
+
+// newTimeTimer creates a new timer using real time.
+func newTimeTimer(d time.Duration) timer {
+	ch := make(chan time.Time)
+	t := time.AfterFunc(d, func() {
+		close(ch)
+	})
+	return &timeTimer{t, ch}
+}
+
+// newTimeAfterFunc creates an AfterFunc timer using real time.
+func newTimeAfterFunc(d time.Duration, f func()) timer {
+	return &timeTimer{
+		t: time.AfterFunc(d, f),
+	}
+}
+
+func (t timeTimer) C() <-chan time.Time        { return t.c }
+func (t timeTimer) Stop() bool                 { return t.t.Stop() }
+func (t timeTimer) Reset(d time.Duration) bool { return t.t.Reset(d) }
+
+// fakeTimer implements timer using fake time.
+type fakeTimer struct {
+	hooks *testSyncHooks
+
+	mu   sync.Mutex
+	when time.Time      // when the timer will fire
+	c    chan time.Time // closed when the timer fires; mutually exclusive with f
+	f    func()         // called when the timer fires; mutually exclusive with c
+}
+
+func (t *fakeTimer) C() <-chan time.Time { return t.c }
+
+func (t *fakeTimer) Stop() bool {
+	t.mu.Lock()
+	defer t.mu.Unlock()
+	stopped := t.when.IsZero()
+	t.when = time.Time{}
+	return stopped
+}
+
+func (t *fakeTimer) Reset(d time.Duration) bool {
+	if t.c != nil || t.f == nil {
+		panic("fakeTimer only supports Reset on AfterFunc timers")
+	}
+	t.mu.Lock()
+	defer t.mu.Unlock()
+	t.hooks.lock()
+	defer t.hooks.unlock()
+	active := !t.when.IsZero()
+	t.when = t.hooks.now.Add(d)
+	if !active {
+		t.hooks.timers = append(t.hooks.timers, t)
+	}
+	return active
+}
diff --git a/vendor/golang.org/x/net/http2/transport.go b/vendor/golang.org/x/net/http2/transport.go
index c2a5b44b3d6..ce375c8c753 100644
--- a/vendor/golang.org/x/net/http2/transport.go
+++ b/vendor/golang.org/x/net/http2/transport.go
@@ -147,6 +147,12 @@ type Transport struct {
 	// waiting for their turn.
 	StrictMaxConcurrentStreams bool
 
+	// IdleConnTimeout is the maximum amount of time an idle
+	// (keep-alive) connection will remain idle before closing
+	// itself.
+	// Zero means no limit.
+	IdleConnTimeout time.Duration
+
 	// ReadIdleTimeout is the timeout after which a health check using ping
 	// frame will be carried out if no frame is received on the connection.
 	// Note that a ping response will is considered a received frame, so if
@@ -178,6 +184,8 @@ type Transport struct {
 
 	connPoolOnce  sync.Once
 	connPoolOrDef ClientConnPool // non-nil version of ConnPool
+
+	syncHooks *testSyncHooks
 }
 
 func (t *Transport) maxHeaderListSize() uint32 {
@@ -302,7 +310,7 @@ type ClientConn struct {
 	readerErr  error         // set before readerDone is closed
 
 	idleTimeout time.Duration // or 0 for never
-	idleTimer   *time.Timer
+	idleTimer   timer
 
 	mu              sync.Mutex // guards following
 	cond            *sync.Cond // hold mu; broadcast on flow/closed changes
@@ -344,6 +352,60 @@ type ClientConn struct {
 	werr error        // first write error that has occurred
 	hbuf bytes.Buffer // HPACK encoder writes into this
 	henc *hpack.Encoder
+
+	syncHooks *testSyncHooks // can be nil
+}
+
+// Hook points used for testing.
+// Outside of tests, cc.syncHooks is nil and these all have minimal implementations.
+// Inside tests, see the testSyncHooks function docs.
+
+// goRun starts a new goroutine.
+func (cc *ClientConn) goRun(f func()) {
+	if cc.syncHooks != nil {
+		cc.syncHooks.goRun(f)
+		return
+	}
+	go f()
+}
+
+// condBroadcast is cc.cond.Broadcast.
+func (cc *ClientConn) condBroadcast() {
+	if cc.syncHooks != nil {
+		cc.syncHooks.condBroadcast(cc.cond)
+	}
+	cc.cond.Broadcast()
+}
+
+// condWait is cc.cond.Wait.
+func (cc *ClientConn) condWait() {
+	if cc.syncHooks != nil {
+		cc.syncHooks.condWait(cc.cond)
+	}
+	cc.cond.Wait()
+}
+
+// newTimer creates a new time.Timer, or a synthetic timer in tests.
+func (cc *ClientConn) newTimer(d time.Duration) timer {
+	if cc.syncHooks != nil {
+		return cc.syncHooks.newTimer(d)
+	}
+	return newTimeTimer(d)
+}
+
+// afterFunc creates a new time.AfterFunc timer, or a synthetic timer in tests.
+func (cc *ClientConn) afterFunc(d time.Duration, f func()) timer {
+	if cc.syncHooks != nil {
+		return cc.syncHooks.afterFunc(d, f)
+	}
+	return newTimeAfterFunc(d, f)
+}
+
+func (cc *ClientConn) contextWithTimeout(ctx context.Context, d time.Duration) (context.Context, context.CancelFunc) {
+	if cc.syncHooks != nil {
+		return cc.syncHooks.contextWithTimeout(ctx, d)
+	}
+	return context.WithTimeout(ctx, d)
 }
 
 // clientStream is the state for a single HTTP/2 stream. One of these
@@ -425,7 +487,7 @@ func (cs *clientStream) abortStreamLocked(err error) {
 	// TODO(dneil): Clean up tests where cs.cc.cond is nil.
 	if cs.cc.cond != nil {
 		// Wake up writeRequestBody if it is waiting on flow control.
-		cs.cc.cond.Broadcast()
+		cs.cc.condBroadcast()
 	}
 }
 
@@ -435,7 +497,7 @@ func (cs *clientStream) abortRequestBodyWrite() {
 	defer cc.mu.Unlock()
 	if cs.reqBody != nil && cs.reqBodyClosed == nil {
 		cs.closeReqBodyLocked()
-		cc.cond.Broadcast()
+		cc.condBroadcast()
 	}
 }
 
@@ -445,10 +507,10 @@ func (cs *clientStream) closeReqBodyLocked() {
 	}
 	cs.reqBodyClosed = make(chan struct{})
 	reqBodyClosed := cs.reqBodyClosed
-	go func() {
+	cs.cc.goRun(func() {
 		cs.reqBody.Close()
 		close(reqBodyClosed)
-	}()
+	})
 }
 
 type stickyErrWriter struct {
@@ -537,15 +599,6 @@ func authorityAddr(scheme string, authority string) (addr string) {
 	return net.JoinHostPort(host, port)
 }
 
-var retryBackoffHook func(time.Duration) *time.Timer
-
-func backoffNewTimer(d time.Duration) *time.Timer {
-	if retryBackoffHook != nil {
-		return retryBackoffHook(d)
-	}
-	return time.NewTimer(d)
-}
-
 // RoundTripOpt is like RoundTrip, but takes options.
 func (t *Transport) RoundTripOpt(req *http.Request, opt RoundTripOpt) (*http.Response, error) {
 	if !(req.URL.Scheme == "https" || (req.URL.Scheme == "http" && t.AllowHTTP)) {
@@ -573,13 +626,27 @@ func (t *Transport) RoundTripOpt(req *http.Request, opt RoundTripOpt) (*http.Res
 				backoff := float64(uint(1) << (uint(retry) - 1))
 				backoff += backoff * (0.1 * mathrand.Float64())
 				d := time.Second * time.Duration(backoff)
-				timer := backoffNewTimer(d)
+				var tm timer
+				if t.syncHooks != nil {
+					tm = t.syncHooks.newTimer(d)
+					t.syncHooks.blockUntil(func() bool {
+						select {
+						case <-tm.C():
+						case <-req.Context().Done():
+						default:
+							return false
+						}
+						return true
+					})
+				} else {
+					tm = newTimeTimer(d)
+				}
 				select {
-				case <-timer.C:
+				case <-tm.C():
 					t.vlogf("RoundTrip retrying after failure: %v", roundTripErr)
 					continue
 				case <-req.Context().Done():
-					timer.Stop()
+					tm.Stop()
 					err = req.Context().Err()
 				}
 			}
@@ -658,6 +725,9 @@ func canRetryError(err error) bool {
 }
 
 func (t *Transport) dialClientConn(ctx context.Context, addr string, singleUse bool) (*ClientConn, error) {
+	if t.syncHooks != nil {
+		return t.newClientConn(nil, singleUse, t.syncHooks)
+	}
 	host, _, err := net.SplitHostPort(addr)
 	if err != nil {
 		return nil, err
@@ -666,7 +736,7 @@ func (t *Transport) dialClientConn(ctx context.Context, addr string, singleUse b
 	if err != nil {
 		return nil, err
 	}
-	return t.newClientConn(tconn, singleUse)
+	return t.newClientConn(tconn, singleUse, nil)
 }
 
 func (t *Transport) newTLSConfig(host string) *tls.Config {
@@ -732,10 +802,10 @@ func (t *Transport) maxEncoderHeaderTableSize() uint32 {
 }
 
 func (t *Transport) NewClientConn(c net.Conn) (*ClientConn, error) {
-	return t.newClientConn(c, t.disableKeepAlives())
+	return t.newClientConn(c, t.disableKeepAlives(), nil)
 }
 
-func (t *Transport) newClientConn(c net.Conn, singleUse bool) (*ClientConn, error) {
+func (t *Transport) newClientConn(c net.Conn, singleUse bool, hooks *testSyncHooks) (*ClientConn, error) {
 	cc := &ClientConn{
 		t:                     t,
 		tconn:                 c,
@@ -750,10 +820,15 @@ func (t *Transport) newClientConn(c net.Conn, singleUse bool) (*ClientConn, erro
 		wantSettingsAck:       true,
 		pings:                 make(map[[8]byte]chan struct{}),
 		reqHeaderMu:           make(chan struct{}, 1),
+		syncHooks:             hooks,
+	}
+	if hooks != nil {
+		hooks.newclientconn(cc)
+		c = cc.tconn
 	}
 	if d := t.idleConnTimeout(); d != 0 {
 		cc.idleTimeout = d
-		cc.idleTimer = time.AfterFunc(d, cc.onIdleTimeout)
+		cc.idleTimer = cc.afterFunc(d, cc.onIdleTimeout)
 	}
 	if VerboseLogs {
 		t.vlogf("http2: Transport creating client conn %p to %v", cc, c.RemoteAddr())
@@ -818,7 +893,7 @@ func (t *Transport) newClientConn(c net.Conn, singleUse bool) (*ClientConn, erro
 		return nil, cc.werr
 	}
 
-	go cc.readLoop()
+	cc.goRun(cc.readLoop)
 	return cc, nil
 }
 
@@ -826,7 +901,7 @@ func (cc *ClientConn) healthCheck() {
 	pingTimeout := cc.t.pingTimeout()
 	// We don't need to periodically ping in the health check, because the readLoop of ClientConn will
 	// trigger the healthCheck again if there is no frame received.
-	ctx, cancel := context.WithTimeout(context.Background(), pingTimeout)
+	ctx, cancel := cc.contextWithTimeout(context.Background(), pingTimeout)
 	defer cancel()
 	cc.vlogf("http2: Transport sending health check")
 	err := cc.Ping(ctx)
@@ -1056,7 +1131,7 @@ func (cc *ClientConn) Shutdown(ctx context.Context) error {
 	// Wait for all in-flight streams to complete or connection to close
 	done := make(chan struct{})
 	cancelled := false // guarded by cc.mu
-	go func() {
+	cc.goRun(func() {
 		cc.mu.Lock()
 		defer cc.mu.Unlock()
 		for {
@@ -1068,9 +1143,9 @@ func (cc *ClientConn) Shutdown(ctx context.Context) error {
 			if cancelled {
 				break
 			}
-			cc.cond.Wait()
+			cc.condWait()
 		}
-	}()
+	})
 	shutdownEnterWaitStateHook()
 	select {
 	case <-done:
@@ -1080,7 +1155,7 @@ func (cc *ClientConn) Shutdown(ctx context.Context) error {
 		cc.mu.Lock()
 		// Free the goroutine above
 		cancelled = true
-		cc.cond.Broadcast()
+		cc.condBroadcast()
 		cc.mu.Unlock()
 		return ctx.Err()
 	}
@@ -1118,7 +1193,7 @@ func (cc *ClientConn) closeForError(err error) {
 	for _, cs := range cc.streams {
 		cs.abortStreamLocked(err)
 	}
-	cc.cond.Broadcast()
+	cc.condBroadcast()
 	cc.mu.Unlock()
 	cc.closeConn()
 }
@@ -1215,6 +1290,10 @@ func (cc *ClientConn) decrStreamReservationsLocked() {
 }
 
 func (cc *ClientConn) RoundTrip(req *http.Request) (*http.Response, error) {
+	return cc.roundTrip(req, nil)
+}
+
+func (cc *ClientConn) roundTrip(req *http.Request, streamf func(*clientStream)) (*http.Response, error) {
 	ctx := req.Context()
 	cs := &clientStream{
 		cc:                   cc,
@@ -1229,9 +1308,23 @@ func (cc *ClientConn) RoundTrip(req *http.Request) (*http.Response, error) {
 		respHeaderRecv:       make(chan struct{}),
 		donec:                make(chan struct{}),
 	}
-	go cs.doRequest(req)
+	cc.goRun(func() {
+		cs.doRequest(req)
+	})
 
 	waitDone := func() error {
+		if cc.syncHooks != nil {
+			cc.syncHooks.blockUntil(func() bool {
+				select {
+				case <-cs.donec:
+				case <-ctx.Done():
+				case <-cs.reqCancel:
+				default:
+					return false
+				}
+				return true
+			})
+		}
 		select {
 		case <-cs.donec:
 			return nil
@@ -1292,7 +1385,24 @@ func (cc *ClientConn) RoundTrip(req *http.Request) (*http.Response, error) {
 		return err
 	}
 
+	if streamf != nil {
+		streamf(cs)
+	}
+
 	for {
+		if cc.syncHooks != nil {
+			cc.syncHooks.blockUntil(func() bool {
+				select {
+				case <-cs.respHeaderRecv:
+				case <-cs.abort:
+				case <-ctx.Done():
+				case <-cs.reqCancel:
+				default:
+					return false
+				}
+				return true
+			})
+		}
 		select {
 		case <-cs.respHeaderRecv:
 			return handleResponseHeaders()
@@ -1348,6 +1458,21 @@ func (cs *clientStream) writeRequest(req *http.Request) (err error) {
 	if cc.reqHeaderMu == nil {
 		panic("RoundTrip on uninitialized ClientConn") // for tests
 	}
+	var newStreamHook func(*clientStream)
+	if cc.syncHooks != nil {
+		newStreamHook = cc.syncHooks.newstream
+		cc.syncHooks.blockUntil(func() bool {
+			select {
+			case cc.reqHeaderMu <- struct{}{}:
+				<-cc.reqHeaderMu
+			case <-cs.reqCancel:
+			case <-ctx.Done():
+			default:
+				return false
+			}
+			return true
+		})
+	}
 	select {
 	case cc.reqHeaderMu <- struct{}{}:
 	case <-cs.reqCancel:
@@ -1372,6 +1497,10 @@ func (cs *clientStream) writeRequest(req *http.Request) (err error) {
 	}
 	cc.mu.Unlock()
 
+	if newStreamHook != nil {
+		newStreamHook(cs)
+	}
+
 	// TODO(bradfitz): this is a copy of the logic in net/http. Unify somewhere?
 	if !cc.t.disableCompression() &&
 		req.Header.Get("Accept-Encoding") == "" &&
@@ -1452,15 +1581,30 @@ func (cs *clientStream) writeRequest(req *http.Request) (err error) {
 	var respHeaderTimer <-chan time.Time
 	var respHeaderRecv chan struct{}
 	if d := cc.responseHeaderTimeout(); d != 0 {
-		timer := time.NewTimer(d)
+		timer := cc.newTimer(d)
 		defer timer.Stop()
-		respHeaderTimer = timer.C
+		respHeaderTimer = timer.C()
 		respHeaderRecv = cs.respHeaderRecv
 	}
 	// Wait until the peer half-closes its end of the stream,
 	// or until the request is aborted (via context, error, or otherwise),
 	// whichever comes first.
 	for {
+		if cc.syncHooks != nil {
+			cc.syncHooks.blockUntil(func() bool {
+				select {
+				case <-cs.peerClosed:
+				case <-respHeaderTimer:
+				case <-respHeaderRecv:
+				case <-cs.abort:
+				case <-ctx.Done():
+				case <-cs.reqCancel:
+				default:
+					return false
+				}
+				return true
+			})
+		}
 		select {
 		case <-cs.peerClosed:
 			return nil
@@ -1609,7 +1753,7 @@ func (cc *ClientConn) awaitOpenSlotForStreamLocked(cs *clientStream) error {
 			return nil
 		}
 		cc.pendingRequests++
-		cc.cond.Wait()
+		cc.condWait()
 		cc.pendingRequests--
 		select {
 		case <-cs.abort:
@@ -1871,8 +2015,24 @@ func (cs *clientStream) awaitFlowControl(maxBytes int) (taken int32, err error)
 			cs.flow.take(take)
 			return take, nil
 		}
-		cc.cond.Wait()
+		cc.condWait()
+	}
+}
+
+func validateHeaders(hdrs http.Header) string {
+	for k, vv := range hdrs {
+		if !httpguts.ValidHeaderFieldName(k) {
+			return fmt.Sprintf("name %q", k)
+		}
+		for _, v := range vv {
+			if !httpguts.ValidHeaderFieldValue(v) {
+				// Don't include the value in the error,
+				// because it may be sensitive.
+				return fmt.Sprintf("value for header %q", k)
+			}
+		}
 	}
+	return ""
 }
 
 var errNilRequestURL = errors.New("http2: Request.URI is nil")
@@ -1912,19 +2072,14 @@ func (cc *ClientConn) encodeHeaders(req *http.Request, addGzipHeader bool, trail
 		}
 	}
 
-	// Check for any invalid headers and return an error before we
+	// Check for any invalid headers+trailers and return an error before we
 	// potentially pollute our hpack state. (We want to be able to
 	// continue to reuse the hpack encoder for future requests)
-	for k, vv := range req.Header {
-		if !httpguts.ValidHeaderFieldName(k) {
-			return nil, fmt.Errorf("invalid HTTP header name %q", k)
-		}
-		for _, v := range vv {
-			if !httpguts.ValidHeaderFieldValue(v) {
-				// Don't include the value in the error, because it may be sensitive.
-				return nil, fmt.Errorf("invalid HTTP header value for header %q", k)
-			}
-		}
+	if err := validateHeaders(req.Header); err != "" {
+		return nil, fmt.Errorf("invalid HTTP header %s", err)
+	}
+	if err := validateHeaders(req.Trailer); err != "" {
+		return nil, fmt.Errorf("invalid HTTP trailer %s", err)
 	}
 
 	enumerateHeaders := func(f func(name, value string)) {
@@ -2143,7 +2298,7 @@ func (cc *ClientConn) forgetStreamID(id uint32) {
 	}
 	// Wake up writeRequestBody via clientStream.awaitFlowControl and
 	// wake up RoundTrip if there is a pending request.
-	cc.cond.Broadcast()
+	cc.condBroadcast()
 
 	closeOnIdle := cc.singleUse || cc.doNotReuse || cc.t.disableKeepAlives() || cc.goAway != nil
 	if closeOnIdle && cc.streamsReserved == 0 && len(cc.streams) == 0 {
@@ -2231,7 +2386,7 @@ func (rl *clientConnReadLoop) cleanup() {
 			cs.abortStreamLocked(err)
 		}
 	}
-	cc.cond.Broadcast()
+	cc.condBroadcast()
 	cc.mu.Unlock()
 }
 
@@ -2266,10 +2421,9 @@ func (rl *clientConnReadLoop) run() error {
 	cc := rl.cc
 	gotSettings := false
 	readIdleTimeout := cc.t.ReadIdleTimeout
-	var t *time.Timer
+	var t timer
 	if readIdleTimeout != 0 {
-		t = time.AfterFunc(readIdleTimeout, cc.healthCheck)
-		defer t.Stop()
+		t = cc.afterFunc(readIdleTimeout, cc.healthCheck)
 	}
 	for {
 		f, err := cc.fr.ReadFrame()
@@ -2684,7 +2838,7 @@ func (rl *clientConnReadLoop) processData(f *DataFrame) error {
 		})
 		return nil
 	}
-	if !cs.firstByte {
+	if !cs.pastHeaders {
 		cc.logf("protocol error: received DATA before a HEADERS frame")
 		rl.endStreamError(cs, StreamError{
 			StreamID: f.StreamID,
@@ -2867,7 +3021,7 @@ func (rl *clientConnReadLoop) processSettingsNoWrite(f *SettingsFrame) error {
 			for _, cs := range cc.streams {
 				cs.flow.add(delta)
 			}
-			cc.cond.Broadcast()
+			cc.condBroadcast()
 
 			cc.initialWindowSize = s.Val
 		case SettingHeaderTableSize:
@@ -2922,7 +3076,7 @@ func (rl *clientConnReadLoop) processWindowUpdate(f *WindowUpdateFrame) error {
 
 		return ConnectionError(ErrCodeFlowControl)
 	}
-	cc.cond.Broadcast()
+	cc.condBroadcast()
 	return nil
 }
 
@@ -2964,24 +3118,38 @@ func (cc *ClientConn) Ping(ctx context.Context) error {
 		}
 		cc.mu.Unlock()
 	}
-	errc := make(chan error, 1)
-	go func() {
+	var pingError error
+	errc := make(chan struct{})
+	cc.goRun(func() {
 		cc.wmu.Lock()
 		defer cc.wmu.Unlock()
-		if err := cc.fr.WritePing(false, p); err != nil {
-			errc <- err
+		if pingError = cc.fr.WritePing(false, p); pingError != nil {
+			close(errc)
 			return
 		}
-		if err := cc.bw.Flush(); err != nil {
-			errc <- err
+		if pingError = cc.bw.Flush(); pingError != nil {
+			close(errc)
 			return
 		}
-	}()
+	})
+	if cc.syncHooks != nil {
+		cc.syncHooks.blockUntil(func() bool {
+			select {
+			case <-c:
+			case <-errc:
+			case <-ctx.Done():
+			case <-cc.readerDone:
+			default:
+				return false
+			}
+			return true
+		})
+	}
 	select {
 	case <-c:
 		return nil
-	case err := <-errc:
-		return err
+	case <-errc:
+		return pingError
 	case <-ctx.Done():
 		return ctx.Err()
 	case <-cc.readerDone:
@@ -3150,9 +3318,17 @@ func (rt noDialH2RoundTripper) RoundTrip(req *http.Request) (*http.Response, err
 }
 
 func (t *Transport) idleConnTimeout() time.Duration {
+	// to keep things backwards compatible, we use non-zero values of
+	// IdleConnTimeout, followed by using the IdleConnTimeout on the underlying
+	// http1 transport, followed by 0
+	if t.IdleConnTimeout != 0 {
+		return t.IdleConnTimeout
+	}
+
 	if t.t1 != nil {
 		return t.t1.IdleConnTimeout
 	}
+
 	return 0
 }
 
diff --git a/vendor/k8s.io/apiserver/pkg/features/kube_features.go b/vendor/k8s.io/apiserver/pkg/features/kube_features.go
index bb944b64f0c..e524e0c6474 100644
--- a/vendor/k8s.io/apiserver/pkg/features/kube_features.go
+++ b/vendor/k8s.io/apiserver/pkg/features/kube_features.go
@@ -240,6 +240,12 @@ const (
 	// Enables support for watch bookmark events.
 	WatchBookmark featuregate.Feature = "WatchBookmark"
 
+	// owner: @serathius
+	// beta: 1.30
+	// Enables watches without resourceVersion to be served from storage.
+	// Used to prevent https://github.com/kubernetes/kubernetes/issues/123072 until etcd fixes the issue.
+	WatchFromStorageWithoutResourceVersion featuregate.Feature = "WatchFromStorageWithoutResourceVersion"
+
 	// owner: @vinaykul
 	// kep: http://kep.k8s.io/1287
 	// alpha: v1.27
@@ -328,6 +334,8 @@ var defaultKubernetesFeatureGates = map[featuregate.Feature]featuregate.FeatureS
 
 	WatchBookmark: {Default: true, PreRelease: featuregate.GA, LockToDefault: true},
 
+	WatchFromStorageWithoutResourceVersion: {Default: false, PreRelease: featuregate.Beta},
+
 	InPlacePodVerticalScaling: {Default: false, PreRelease: featuregate.Alpha},
 
 	WatchList: {Default: false, PreRelease: featuregate.Alpha},
diff --git a/vendor/modules.txt b/vendor/modules.txt
index 5b9bdbc748e..6cea0a33e91 100644
--- a/vendor/modules.txt
+++ b/vendor/modules.txt
@@ -730,7 +730,7 @@ golang.org/x/crypto/internal/poly1305
 golang.org/x/exp/constraints
 golang.org/x/exp/maps
 golang.org/x/exp/slices
-# golang.org/x/net v0.22.0
+# golang.org/x/net v0.23.0
 ## explicit; go 1.18
 golang.org/x/net/context
 golang.org/x/net/html
@@ -971,7 +971,7 @@ gopkg.in/yaml.v2
 # gopkg.in/yaml.v3 v3.0.1
 ## explicit
 gopkg.in/yaml.v3
-# k8s.io/api v0.29.3
+# k8s.io/api v0.29.4
 ## explicit; go 1.21
 k8s.io/api/admission/v1
 k8s.io/api/admission/v1beta1
@@ -1027,7 +1027,7 @@ k8s.io/api/scheduling/v1beta1
 k8s.io/api/storage/v1
 k8s.io/api/storage/v1alpha1
 k8s.io/api/storage/v1beta1
-# k8s.io/apiextensions-apiserver v0.29.3
+# k8s.io/apiextensions-apiserver v0.29.4
 ## explicit; go 1.21
 k8s.io/apiextensions-apiserver/pkg/apihelpers
 k8s.io/apiextensions-apiserver/pkg/apis/apiextensions
@@ -1048,7 +1048,7 @@ k8s.io/apiextensions-apiserver/pkg/client/clientset/clientset/scheme
 k8s.io/apiextensions-apiserver/pkg/client/clientset/clientset/typed/apiextensions/v1
 k8s.io/apiextensions-apiserver/pkg/client/clientset/clientset/typed/apiextensions/v1beta1
 k8s.io/apiextensions-apiserver/pkg/features
-# k8s.io/apimachinery v0.29.3
+# k8s.io/apimachinery v0.29.4
 ## explicit; go 1.21
 k8s.io/apimachinery/pkg/api/equality
 k8s.io/apimachinery/pkg/api/errors
@@ -1102,7 +1102,7 @@ k8s.io/apimachinery/pkg/version
 k8s.io/apimachinery/pkg/watch
 k8s.io/apimachinery/third_party/forked/golang/json
 k8s.io/apimachinery/third_party/forked/golang/reflect
-# k8s.io/apiserver v0.29.3
+# k8s.io/apiserver v0.29.4
 ## explicit; go 1.21
 k8s.io/apiserver/pkg/admission
 k8s.io/apiserver/pkg/admission/cel
@@ -1146,7 +1146,7 @@ k8s.io/apiserver/pkg/util/feature
 k8s.io/apiserver/pkg/util/webhook
 k8s.io/apiserver/pkg/util/x509metrics
 k8s.io/apiserver/pkg/warning
-# k8s.io/client-go v0.29.3
+# k8s.io/client-go v0.29.4
 ## explicit; go 1.21
 k8s.io/client-go/applyconfigurations/admissionregistration/v1
 k8s.io/client-go/applyconfigurations/admissionregistration/v1alpha1
@@ -1402,7 +1402,7 @@ k8s.io/client-go/util/homedir
 k8s.io/client-go/util/keyutil
 k8s.io/client-go/util/retry
 k8s.io/client-go/util/workqueue
-# k8s.io/component-base v0.29.3
+# k8s.io/component-base v0.29.4
 ## explicit; go 1.21
 k8s.io/component-base/config
 k8s.io/component-base/config/v1alpha1
@@ -1468,7 +1468,7 @@ sigs.k8s.io/apiserver-network-proxy/konnectivity-client/pkg/client
 sigs.k8s.io/apiserver-network-proxy/konnectivity-client/pkg/client/metrics
 sigs.k8s.io/apiserver-network-proxy/konnectivity-client/pkg/common/metrics
 sigs.k8s.io/apiserver-network-proxy/konnectivity-client/proto/client
-# sigs.k8s.io/controller-runtime v0.17.2
+# sigs.k8s.io/controller-runtime v0.17.5
 ## explicit; go 1.21
 sigs.k8s.io/controller-runtime
 sigs.k8s.io/controller-runtime/pkg/builder
diff --git a/vendor/sigs.k8s.io/controller-runtime/pkg/cache/cache.go b/vendor/sigs.k8s.io/controller-runtime/pkg/cache/cache.go
index 1cecf88e5e5..73ad68fe43a 100644
--- a/vendor/sigs.k8s.io/controller-runtime/pkg/cache/cache.go
+++ b/vendor/sigs.k8s.io/controller-runtime/pkg/cache/cache.go
@@ -20,6 +20,7 @@ import (
 	"context"
 	"fmt"
 	"net/http"
+	"sort"
 	"time"
 
 	"golang.org/x/exp/maps"
@@ -418,14 +419,6 @@ func defaultOpts(config *rest.Config, opts Options) (Options, error) {
 		}
 	}
 
-	for namespace, cfg := range opts.DefaultNamespaces {
-		cfg = defaultConfig(cfg, optionDefaultsToConfig(&opts))
-		if namespace == metav1.NamespaceAll {
-			cfg.FieldSelector = fields.AndSelectors(appendIfNotNil(namespaceAllSelector(maps.Keys(opts.DefaultNamespaces)), cfg.FieldSelector)...)
-		}
-		opts.DefaultNamespaces[namespace] = cfg
-	}
-
 	for obj, byObject := range opts.ByObject {
 		isNamespaced, err := apiutil.IsObjectNamespaced(obj, opts.Scheme, opts.Mapper)
 		if err != nil {
@@ -435,7 +428,12 @@ func defaultOpts(config *rest.Config, opts Options) (Options, error) {
 			return opts, fmt.Errorf("type %T is not namespaced, but its ByObject.Namespaces setting is not nil", obj)
 		}
 
-		// Default the namespace-level configs first, because they need to use the undefaulted type-level config.
+		if isNamespaced && byObject.Namespaces == nil {
+			byObject.Namespaces = maps.Clone(opts.DefaultNamespaces)
+		}
+
+		// Default the namespace-level configs first, because they need to use the undefaulted type-level config
+		// to be able to potentially fall through to settings from DefaultNamespaces.
 		for namespace, config := range byObject.Namespaces {
 			// 1. Default from the undefaulted type-level config
 			config = defaultConfig(config, byObjectToConfig(byObject))
@@ -461,19 +459,35 @@ func defaultOpts(config *rest.Config, opts Options) (Options, error) {
 			byObject.Namespaces[namespace] = config
 		}
 
-		defaultedConfig := defaultConfig(byObjectToConfig(byObject), optionDefaultsToConfig(&opts))
-		byObject.Label = defaultedConfig.LabelSelector
-		byObject.Field = defaultedConfig.FieldSelector
-		byObject.Transform = defaultedConfig.Transform
-		byObject.UnsafeDisableDeepCopy = defaultedConfig.UnsafeDisableDeepCopy
-
-		if isNamespaced && byObject.Namespaces == nil {
-			byObject.Namespaces = opts.DefaultNamespaces
+		// Only default ByObject iself if it isn't namespaced or has no namespaces configured, as only
+		// then any of this will be honored.
+		if !isNamespaced || len(byObject.Namespaces) == 0 {
+			defaultedConfig := defaultConfig(byObjectToConfig(byObject), optionDefaultsToConfig(&opts))
+			byObject.Label = defaultedConfig.LabelSelector
+			byObject.Field = defaultedConfig.FieldSelector
+			byObject.Transform = defaultedConfig.Transform
+			byObject.UnsafeDisableDeepCopy = defaultedConfig.UnsafeDisableDeepCopy
 		}
 
 		opts.ByObject[obj] = byObject
 	}
 
+	// Default namespaces after byObject has been defaulted, otherwise a namespace without selectors
+	// will get the `Default` selectors, then get copied to byObject and then not get defaulted from
+	// byObject, as it already has selectors.
+	for namespace, cfg := range opts.DefaultNamespaces {
+		cfg = defaultConfig(cfg, optionDefaultsToConfig(&opts))
+		if namespace == metav1.NamespaceAll {
+			cfg.FieldSelector = fields.AndSelectors(
+				appendIfNotNil(
+					namespaceAllSelector(maps.Keys(opts.DefaultNamespaces)),
+					cfg.FieldSelector,
+				)...,
+			)
+		}
+		opts.DefaultNamespaces[namespace] = cfg
+	}
+
 	// Default the resync period to 10 hours if unset
 	if opts.SyncPeriod == nil {
 		opts.SyncPeriod = &defaultSyncPeriod
@@ -498,20 +512,21 @@ func defaultConfig(toDefault, defaultFrom Config) Config {
 	return toDefault
 }
 
-func namespaceAllSelector(namespaces []string) fields.Selector {
+func namespaceAllSelector(namespaces []string) []fields.Selector {
 	selectors := make([]fields.Selector, 0, len(namespaces)-1)
+	sort.Strings(namespaces)
 	for _, namespace := range namespaces {
 		if namespace != metav1.NamespaceAll {
 			selectors = append(selectors, fields.OneTermNotEqualSelector("metadata.namespace", namespace))
 		}
 	}
 
-	return fields.AndSelectors(selectors...)
+	return selectors
 }
 
-func appendIfNotNil[T comparable](a, b T) []T {
+func appendIfNotNil[T comparable](a []T, b T) []T {
 	if b != *new(T) {
-		return []T{a, b}
+		return append(a, b)
 	}
-	return []T{a}
+	return a
 }
diff --git a/vendor/sigs.k8s.io/controller-runtime/pkg/client/fieldowner.go b/vendor/sigs.k8s.io/controller-runtime/pkg/client/fieldowner.go
new file mode 100644
index 00000000000..2f2f892ef3f
--- /dev/null
+++ b/vendor/sigs.k8s.io/controller-runtime/pkg/client/fieldowner.go
@@ -0,0 +1,106 @@
+/*
+Copyright 2024 The Kubernetes 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 client
+
+import (
+	"context"
+
+	"k8s.io/apimachinery/pkg/api/meta"
+	"k8s.io/apimachinery/pkg/runtime"
+	"k8s.io/apimachinery/pkg/runtime/schema"
+)
+
+// WithFieldOwner wraps a Client and adds the fieldOwner as the field
+// manager to all write requests from this client. If additional [FieldOwner]
+// options are specified on methods of this client, the value specified here
+// will be overridden.
+func WithFieldOwner(c Client, fieldOwner string) Client {
+	return &clientWithFieldManager{
+		manager: fieldOwner,
+		c:       c,
+		Reader:  c,
+	}
+}
+
+type clientWithFieldManager struct {
+	manager string
+	c       Client
+	Reader
+}
+
+func (f *clientWithFieldManager) Create(ctx context.Context, obj Object, opts ...CreateOption) error {
+	return f.c.Create(ctx, obj, append([]CreateOption{FieldOwner(f.manager)}, opts...)...)
+}
+
+func (f *clientWithFieldManager) Update(ctx context.Context, obj Object, opts ...UpdateOption) error {
+	return f.c.Update(ctx, obj, append([]UpdateOption{FieldOwner(f.manager)}, opts...)...)
+}
+
+func (f *clientWithFieldManager) Patch(ctx context.Context, obj Object, patch Patch, opts ...PatchOption) error {
+	return f.c.Patch(ctx, obj, patch, append([]PatchOption{FieldOwner(f.manager)}, opts...)...)
+}
+
+func (f *clientWithFieldManager) Delete(ctx context.Context, obj Object, opts ...DeleteOption) error {
+	return f.c.Delete(ctx, obj, opts...)
+}
+
+func (f *clientWithFieldManager) DeleteAllOf(ctx context.Context, obj Object, opts ...DeleteAllOfOption) error {
+	return f.c.DeleteAllOf(ctx, obj, opts...)
+}
+
+func (f *clientWithFieldManager) Scheme() *runtime.Scheme     { return f.c.Scheme() }
+func (f *clientWithFieldManager) RESTMapper() meta.RESTMapper { return f.c.RESTMapper() }
+func (f *clientWithFieldManager) GroupVersionKindFor(obj runtime.Object) (schema.GroupVersionKind, error) {
+	return f.c.GroupVersionKindFor(obj)
+}
+func (f *clientWithFieldManager) IsObjectNamespaced(obj runtime.Object) (bool, error) {
+	return f.c.IsObjectNamespaced(obj)
+}
+
+func (f *clientWithFieldManager) Status() StatusWriter {
+	return &subresourceClientWithFieldOwner{
+		owner:             f.manager,
+		subresourceWriter: f.c.Status(),
+	}
+}
+
+func (f *clientWithFieldManager) SubResource(subresource string) SubResourceClient {
+	c := f.c.SubResource(subresource)
+	return &subresourceClientWithFieldOwner{
+		owner:             f.manager,
+		subresourceWriter: c,
+		SubResourceReader: c,
+	}
+}
+
+type subresourceClientWithFieldOwner struct {
+	owner             string
+	subresourceWriter SubResourceWriter
+	SubResourceReader
+}
+
+func (f *subresourceClientWithFieldOwner) Create(ctx context.Context, obj Object, subresource Object, opts ...SubResourceCreateOption) error {
+	return f.subresourceWriter.Create(ctx, obj, subresource, append([]SubResourceCreateOption{FieldOwner(f.owner)}, opts...)...)
+}
+
+func (f *subresourceClientWithFieldOwner) Update(ctx context.Context, obj Object, opts ...SubResourceUpdateOption) error {
+	return f.subresourceWriter.Update(ctx, obj, append([]SubResourceUpdateOption{FieldOwner(f.owner)}, opts...)...)
+}
+
+func (f *subresourceClientWithFieldOwner) Patch(ctx context.Context, obj Object, patch Patch, opts ...SubResourcePatchOption) error {
+	return f.subresourceWriter.Patch(ctx, obj, patch, append([]SubResourcePatchOption{FieldOwner(f.owner)}, opts...)...)
+}
diff --git a/vendor/sigs.k8s.io/controller-runtime/pkg/manager/internal.go b/vendor/sigs.k8s.io/controller-runtime/pkg/manager/internal.go
index a16f354a1b1..2b03263de80 100644
--- a/vendor/sigs.k8s.io/controller-runtime/pkg/manager/internal.go
+++ b/vendor/sigs.k8s.io/controller-runtime/pkg/manager/internal.go
@@ -179,6 +179,24 @@ func (cm *controllerManager) add(r Runnable) error {
 	return cm.runnables.Add(r)
 }
 
+// AddMetricsServerExtraHandler adds extra handler served on path to the http server that serves metrics.
+func (cm *controllerManager) AddMetricsServerExtraHandler(path string, handler http.Handler) error {
+	cm.Lock()
+	defer cm.Unlock()
+	if cm.started {
+		return fmt.Errorf("unable to add new metrics handler because metrics endpoint has already been created")
+	}
+	if cm.metricsServer == nil {
+		cm.GetLogger().Info("warn: metrics server is currently disabled, registering extra handler %q will be ignored", path)
+		return nil
+	}
+	if err := cm.metricsServer.AddExtraHandler(path, handler); err != nil {
+		return err
+	}
+	cm.logger.V(2).Info("Registering metrics http server extra handler", "path", path)
+	return nil
+}
+
 // AddHealthzCheck allows you to add Healthz checker.
 func (cm *controllerManager) AddHealthzCheck(name string, check healthz.Checker) error {
 	cm.Lock()
@@ -518,6 +536,8 @@ func (cm *controllerManager) engageStopProcedure(stopComplete <-chan struct{}) e
 
 		// Stop all the leader election runnables, which includes reconcilers.
 		cm.logger.Info("Stopping and waiting for leader election runnables")
+		// Prevent leader election when shutting down a non-elected manager
+		cm.runnables.LeaderElection.startOnce.Do(func() {})
 		cm.runnables.LeaderElection.StopAndWait(cm.shutdownCtx)
 
 		// Stop the caches before the leader election runnables, this is an important
diff --git a/vendor/sigs.k8s.io/controller-runtime/pkg/manager/manager.go b/vendor/sigs.k8s.io/controller-runtime/pkg/manager/manager.go
index 25c3c7375b2..0b7a865004a 100644
--- a/vendor/sigs.k8s.io/controller-runtime/pkg/manager/manager.go
+++ b/vendor/sigs.k8s.io/controller-runtime/pkg/manager/manager.go
@@ -67,6 +67,15 @@ type Manager interface {
 	// election was configured.
 	Elected() <-chan struct{}
 
+	// AddMetricsServerExtraHandler adds an extra handler served on path to the http server that serves metrics.
+	// Might be useful to register some diagnostic endpoints e.g. pprof.
+	//
+	// Note that these endpoints are meant to be sensitive and shouldn't be exposed publicly.
+	//
+	// If the simple path -> handler mapping offered here is not enough,
+	// a new http server/listener should be added as Runnable to the manager via Add method.
+	AddMetricsServerExtraHandler(path string, handler http.Handler) error
+
 	// AddHealthzCheck allows you to add Healthz checker
 	AddHealthzCheck(name string, check healthz.Checker) error
 
diff --git a/vendor/sigs.k8s.io/controller-runtime/pkg/manager/runnable_group.go b/vendor/sigs.k8s.io/controller-runtime/pkg/manager/runnable_group.go
index 96566f5df11..60609104857 100644
--- a/vendor/sigs.k8s.io/controller-runtime/pkg/manager/runnable_group.go
+++ b/vendor/sigs.k8s.io/controller-runtime/pkg/manager/runnable_group.go
@@ -263,6 +263,15 @@ func (r *runnableGroup) Add(rn Runnable, ready runnableCheck) error {
 		r.start.Unlock()
 	}
 
+	// Recheck if we're stopped and hold the readlock, given that the stop and start can be called
+	// at the same time, we can end up in a situation where the runnable is added
+	// after the group is stopped and the channel is closed.
+	r.stop.RLock()
+	defer r.stop.RUnlock()
+	if r.stopped {
+		return errRunnableGroupStopped
+	}
+
 	// Enqueue the runnable.
 	r.ch <- readyRunnable
 	return nil
@@ -272,7 +281,11 @@ func (r *runnableGroup) Add(rn Runnable, ready runnableCheck) error {
 func (r *runnableGroup) StopAndWait(ctx context.Context) {
 	r.stopOnce.Do(func() {
 		// Close the reconciler channel once we're done.
-		defer close(r.ch)
+		defer func() {
+			r.stop.Lock()
+			close(r.ch)
+			r.stop.Unlock()
+		}()
 
 		_ = r.Start(ctx)
 		r.stop.Lock()
diff --git a/vendor/sigs.k8s.io/controller-runtime/pkg/metrics/server/server.go b/vendor/sigs.k8s.io/controller-runtime/pkg/metrics/server/server.go
index e10c5c2103f..40eb9db8cc5 100644
--- a/vendor/sigs.k8s.io/controller-runtime/pkg/metrics/server/server.go
+++ b/vendor/sigs.k8s.io/controller-runtime/pkg/metrics/server/server.go
@@ -46,6 +46,9 @@ var DefaultBindAddress = ":8080"
 
 // Server is a server that serves metrics.
 type Server interface {
+	// AddExtraHandler adds extra handler served on path to the http server that serves metrics.
+	AddExtraHandler(path string, handler http.Handler) error
+
 	// NeedLeaderElection implements the LeaderElectionRunnable interface, which indicates
 	// the metrics server doesn't need leader election.
 	NeedLeaderElection() bool
@@ -179,6 +182,23 @@ func (*defaultServer) NeedLeaderElection() bool {
 	return false
 }
 
+// AddExtraHandler adds extra handler served on path to the http server that serves metrics.
+func (s *defaultServer) AddExtraHandler(path string, handler http.Handler) error {
+	s.mu.Lock()
+	defer s.mu.Unlock()
+	if s.options.ExtraHandlers == nil {
+		s.options.ExtraHandlers = make(map[string]http.Handler)
+	}
+	if path == defaultMetricsEndpoint {
+		return fmt.Errorf("overriding builtin %s endpoint is not allowed", defaultMetricsEndpoint)
+	}
+	if _, found := s.options.ExtraHandlers[path]; found {
+		return fmt.Errorf("can't register extra handler by duplicate path %q on metrics http server", path)
+	}
+	s.options.ExtraHandlers[path] = handler
+	return nil
+}
+
 // Start runs the server.
 // It will install the metrics related resources depend on the server configuration.
 func (s *defaultServer) Start(ctx context.Context) error {