diff --git a/pkg/util/ring/adaptive_chan.go b/pkg/util/ring/adaptive_chan.go
new file mode 100644
index 00000000000..3087d176a4a
--- /dev/null
+++ b/pkg/util/ring/adaptive_chan.go
@@ -0,0 +1,285 @@
+package ring
+
+import (
+	"context"
+	"errors"
+	"sync"
+)
+
+// Package level named errors.
+var (
+	ErrAdaptiveChanClosed           = errors.New("closed AdaptiveChan")
+	ErrAdaptiveChanControllerClosed = errors.New("closed AdaptiveChanController")
+)
+
+// AdaptiveChan provides a queueing system based on a send-only, a receive-only,
+// and an internal ring buffer queue backed by a *Ring. It also provides an
+// AdaptiveChanController to provide stats and some control on the internal
+// *Ring. Termination is controlled by closing the returned send-only channel.
+// After doing so, the receive-only channel will have the chance to receive all
+// the items still in the queue and will be immediately closed afterwards. Once
+// both channels are closed, the AdaptiveChanController will no longer be usable
+// and will only return ErrAdaptiveChanClosed for all its methods. Leaving the
+// growth and shrinkage of the internal *Ring apart, which can be controlled
+// with the AdaptiveChanController, the implementation is allocation free.
+//
+// The implementation explicitly returns two channels and a struct, instead of
+// just one struct that has the channels, to make a clear statement about the
+// intended usage pattern:
+//
+//  1. Create an adaptive channel.
+//  2. Provide the send-only channel to your producer(s). They are responsible
+//     for closing this channel when they're done. If more than one goroutine
+//     will have access to this channel, then it's the producer's responsibility
+//     to coordinate the channel close operation.
+//  3. Provide the receive-only channel to your consumer(s), and let them
+//     receive from it with the two return value syntax for channels in order to
+//     check for termination from the sending side.
+//  4. Use the AdaptiveChanController to control the internal buffer behaviour
+//     and to monitor stats. This should typically be held by the creator of the
+//     adaptive channel. Refrain from holding a reference to the send-only
+//     channel to force termination of the producing side. Instead, provide a
+//     side mechanism to communicate the intention of terminating the sending
+//     side, e.g. providing your producer(s) with a context as well as the
+//     send-only channel. An adaptive channel is meant as a queueing system, not
+//     as a coordination mechanism for producer(s), consumer(s) and
+//     controller(s).
+//
+// This pattern is designed to maximize decoupling while providing insights and
+// granular control on memory usage. While the controller is not meant to make
+// any direct changes to the queued data, the Clear method provides the
+// opportunity to discard all queued items as an administrative measure. This
+// doesn't terminate the queue, though, i.e. it doesn't close the send-only
+// channel.
+func AdaptiveChan[T any]() (send chan<- T, recv <-chan T, ctrl *AdaptiveChanController) {
+	internalSend := make(chan T)
+	internalRecv := make(chan T)
+	statsChan := make(chan AdaptiveChanStats)
+	cmdChan := make(chan acCmd)
+	ctrl = &AdaptiveChanController{
+		statsChan: statsChan,
+		cmdChan:   cmdChan,
+	}
+
+	go func() {
+		defer close(internalRecv)
+		defer close(statsChan)
+
+		var q Ring[T]
+		var stats AdaptiveChanStats
+
+		// the loop condition is that we either have items to dequeue or that we
+		// have the possibility to receive new items to be queued
+		for q.Len() > 0 || internalSend != nil {
+			// NOTE: the overhead of writing stats in each iteration is
+			// negligible. I tried a two phase stats writing with a chan
+			// struct{} to get notified that the controller wanted stats, then
+			// updating the stats and finally writing to statsChan. There was no
+			// observable difference for just enqueueing and dequeueing after
+			// running the benchmarks several times, and reading stats got worse
+			// by ~22%
+			q.WriteStats(&stats.RingStats)
+
+			// if we don't have anything in the queue, then make the dequeueing
+			// branch of the select block indefinitely by providing a nil
+			// channel, leaving only the queueing branch available
+			dequeueChan := internalRecv
+			if q.Len() == 0 {
+				dequeueChan = nil
+			}
+
+			select {
+			case v, ok := <-internalSend: // blocks until something is queued
+				if !ok {
+					// in was closed, so if we leave it like that the next
+					// iteration will keep receiving zero values with ok=false
+					// without any blocking. So we set in to nil, so that
+					// the next iteration the select will block indefinitely on
+					// this branch of the select and leave only the dequeing
+					// branch active until all items have been dequeued
+					internalSend = nil
+				} else {
+					q.Enqueue(v)
+				}
+
+			case dequeueChan <- q.Peek(): // blocks if nothing to dequeue
+				// we don't want to call Dequeue in the `case` above since that
+				// would consume the item and it would be lost if the queueing
+				// branch was selected, so instead we Peek in the `case` and we
+				// do the actual dequeueing here once we know this branch was
+				// selected
+				q.Dequeue()
+
+			case statsChan <- stats:
+				// stats reading
+				stats.StatsRead++
+
+			case cmd, ok := <-cmdChan:
+				if !ok {
+					// AdaptiveChanController was closed. Set cmdChan to nil so
+					// this branch blocks in the next iteration
+					cmdChan = nil
+					continue
+				}
+
+				// execute a command on the internal *Ring
+				switch cmd.acCmdType {
+				case acCmdMin:
+					q.Min = cmd.intValue
+					stats.Min = cmd.intValue
+				case acCmdMax:
+					q.Max = cmd.intValue
+					stats.Max = cmd.intValue
+				case acCmdGrow:
+					q.Grow(cmd.intValue)
+				case acCmdShrink:
+					q.Shrink(cmd.intValue)
+				case acCmdClear:
+					q.Clear()
+				}
+			}
+		}
+	}()
+
+	return internalSend, internalRecv, ctrl
+}
+
+type acCmdType uint8
+
+const (
+	acCmdMin = iota
+	acCmdMax
+	acCmdClear
+	acCmdGrow
+	acCmdShrink
+)
+
+type acCmd struct {
+	acCmdType
+	intValue int
+}
+
+// AdaptiveChanController provides access to an AdaptiveChan's internal *Ring.
+type AdaptiveChanController struct {
+	statsChan <-chan AdaptiveChanStats
+	cmdChan   chan<- acCmd
+	cmdChanMu sync.Mutex
+}
+
+// Close releases resources associated with this controller. After calling this
+// method, all other methods will return ErrAdaptiveChanControllerClosed. It is
+// idempotent. This doesn't affect the queue itself, but rather prevents further
+// administrative tasks to be performed through the AdaptiveChanController.
+func (r *AdaptiveChanController) Close() {
+	r.cmdChanMu.Lock()
+	defer r.cmdChanMu.Unlock()
+
+	if r.cmdChan != nil {
+		close(r.cmdChan)
+		r.cmdChan = nil
+	}
+}
+
+// Min sets the value of Min in the internal *Ring.
+func (r *AdaptiveChanController) Min(ctx context.Context, n int) error {
+	r.cmdChanMu.Lock()
+	defer r.cmdChanMu.Unlock()
+
+	return sendOrErr(ctx, r.cmdChan, acCmd{
+		acCmdType: acCmdMin,
+		intValue:  n,
+	})
+}
+
+// Max sets the value of Max in the internal *Ring.
+func (r *AdaptiveChanController) Max(ctx context.Context, n int) error {
+	r.cmdChanMu.Lock()
+	defer r.cmdChanMu.Unlock()
+
+	return sendOrErr(ctx, r.cmdChan, acCmd{
+		acCmdType: acCmdMax,
+		intValue:  n,
+	})
+}
+
+// Grow calls Grow on the internal *Ring.
+func (r *AdaptiveChanController) Grow(ctx context.Context, n int) error {
+	r.cmdChanMu.Lock()
+	defer r.cmdChanMu.Unlock()
+
+	return sendOrErr(ctx, r.cmdChan, acCmd{
+		acCmdType: acCmdGrow,
+		intValue:  n,
+	})
+}
+
+// Shrink calls Shrink on the internal *Ring.
+func (r *AdaptiveChanController) Shrink(ctx context.Context, n int) error {
+	r.cmdChanMu.Lock()
+	defer r.cmdChanMu.Unlock()
+
+	return sendOrErr(ctx, r.cmdChan, acCmd{
+		acCmdType: acCmdShrink,
+		intValue:  n,
+	})
+}
+
+// Clear calls Clear on the internal *Ring.
+func (r *AdaptiveChanController) Clear(ctx context.Context) error {
+	r.cmdChanMu.Lock()
+	defer r.cmdChanMu.Unlock()
+
+	return sendOrErr(ctx, r.cmdChan, acCmd{
+		acCmdType: acCmdClear,
+	})
+}
+
+// WriteStats writes a snapshot of general stats about the associated
+// AdaptiveChan to the given *AdaptiveChanStats.
+func (r *AdaptiveChanController) WriteStats(ctx context.Context, s *AdaptiveChanStats) error {
+	r.cmdChanMu.Lock()
+	defer r.cmdChanMu.Unlock()
+	if r.cmdChan == nil {
+		return ErrAdaptiveChanControllerClosed
+	}
+
+	return recvOrErr(ctx, r.statsChan, s)
+}
+
+// AdaptiveChanStats is a snapshot of general stats for an AdaptiveChan.
+type AdaptiveChanStats struct {
+	RingStats
+	// Min is the value of Min in the internal *Ring.
+	Min int
+	// Max value of Max in the internal *Ring.
+	Max int
+	// StatsRead is the total number of stats read before this snapshot. If it
+	// is zero, it means this snapshot is the first reading.
+	StatsRead uint64
+}
+
+func sendOrErr[T any](ctx context.Context, c chan<- T, v T) error {
+	if c == nil {
+		return ErrAdaptiveChanControllerClosed
+	}
+
+	select {
+	case c <- v:
+		return nil
+	case <-ctx.Done():
+		return ctx.Err()
+	}
+}
+
+func recvOrErr[T any](ctx context.Context, c <-chan T, tptr *T) error {
+	select {
+	case t, ok := <-c:
+		if !ok {
+			return ErrAdaptiveChanClosed
+		}
+		*tptr = t
+		return nil
+	case <-ctx.Done():
+		return ctx.Err()
+	}
+}
diff --git a/pkg/util/ring/adaptive_chan_bench_test.go b/pkg/util/ring/adaptive_chan_bench_test.go
new file mode 100644
index 00000000000..6e07e43beaf
--- /dev/null
+++ b/pkg/util/ring/adaptive_chan_bench_test.go
@@ -0,0 +1,73 @@
+package ring
+
+import (
+	"context"
+	"testing"
+)
+
+func BenchmarkAdaptiveChanBaseline(b *testing.B) {
+	in, out, _ := AdaptiveChan[int]()
+	in <- 1
+	<-out
+	b.Cleanup(func() {
+		close(in)
+	})
+
+	b.ResetTimer()
+	b.ReportAllocs()
+
+	for i := 0; i < b.N; i++ {
+		in <- i
+		val := <-out
+		if val != i {
+			b.Fatalf("expected 1, got %d", val)
+		}
+	}
+}
+
+func BenchmarkAdaptiveChanWithStatsRead(b *testing.B) {
+	var stats AdaptiveChanStats
+	in, out, sr := AdaptiveChan[int]()
+	in <- 1
+	<-out
+	ctx := context.Background()
+	b.Cleanup(func() {
+		close(in)
+	})
+
+	b.ResetTimer()
+	b.ReportAllocs()
+
+	for i := 0; i < b.N; i++ {
+		in <- 1
+		val := <-out
+		if val != 1 {
+			b.Fatalf("expected 1, got %d", val)
+		}
+		err := sr.WriteStats(ctx, &stats)
+		if err != nil {
+			b.Fatalf("unexpected error: %v", err)
+		}
+		if stats.Enqueued == 0 {
+			b.Fatalf("unexpected stats: %v", stats)
+		}
+	}
+}
+
+func BenchmarkGoChanBaseline(b *testing.B) {
+	c := make(chan int, 1)
+	b.Cleanup(func() {
+		close(c)
+	})
+
+	b.ResetTimer()
+	b.ReportAllocs()
+
+	for i := 0; i < b.N; i++ {
+		c <- 1
+		val := <-c
+		if val != 1 {
+			b.Fatalf("expected 1, got %d", val)
+		}
+	}
+}
diff --git a/pkg/util/ring/adaptive_chan_test.go b/pkg/util/ring/adaptive_chan_test.go
new file mode 100644
index 00000000000..203d54c0b09
--- /dev/null
+++ b/pkg/util/ring/adaptive_chan_test.go
@@ -0,0 +1,272 @@
+package ring
+
+import (
+	"context"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/require"
+	"go.uber.org/goleak"
+)
+
+func TestMain(m *testing.M) {
+	// make sure we don't leak goroutines after tests in this package have
+	// finished. This is especially important as AdaptiveChan uses a different
+	// goroutine to coordinate work
+	goleak.VerifyTestMain(m)
+}
+
+func TestAdaptiveChan(t *testing.T) {
+	t.Parallel()
+
+	t.Run("edge case - close send and controller after creation", func(t *testing.T) {
+		t.Parallel()
+		send, recv, ctrl := AdaptiveChan[int]()
+		cleanupAC(t, send, recv, ctrl)
+	})
+
+	t.Run("basic operation", func(t *testing.T) {
+		t.Parallel()
+		var stats, expectedStats AdaptiveChanStats
+
+		send, recv, ctrl := AdaptiveChan[int]()
+		cleanupAC(t, send, recv, ctrl)
+
+		sendNonBlock(t, send, ints(10)...)
+		err := ctrl.WriteStats(ctxFromTest(t), &stats)
+		require.NoError(t, err)
+		removeAllocStats(&stats.RingStats)
+		expectedStats.Len = 10
+		expectedStats.Enqueued = 10
+		require.Equal(t, expectedStats, stats)
+
+		recvNonBlock(t, recv, ints(10)...)
+		err = ctrl.WriteStats(ctxFromTest(t), &stats)
+		require.NoError(t, err)
+		removeAllocStats(&stats.RingStats)
+		expectedStats.Len = 0
+		expectedStats.Dequeued = 10
+		expectedStats.StatsRead = 1
+		require.Equal(t, expectedStats, stats)
+	})
+
+	t.Run("using commands to control the ring", func(t *testing.T) {
+		t.Parallel()
+		send, recv, ctrl := AdaptiveChan[int]()
+		cleanupAC(t, send, recv, ctrl)
+
+		var stats, expectedStats AdaptiveChanStats
+		expectedStats.Min = 10
+		expectedStats.Max = 20
+
+		err := ctrl.Min(ctxFromTest(t), expectedStats.Min)
+		require.NoError(t, err)
+		err = ctrl.Max(ctxFromTest(t), expectedStats.Max)
+		require.NoError(t, err)
+
+		sendNonBlock(t, send, 1)
+		err = ctrl.WriteStats(ctxFromTest(t), &stats)
+		require.NoError(t, err)
+		require.Equal(t, expectedStats.Min, stats.Cap, "failed to allocate Min")
+		removeAllocStats(&stats.RingStats)
+		expectedStats.Len = 1
+		expectedStats.Enqueued = 1
+		require.Equal(t, expectedStats, stats)
+
+		err = ctrl.Grow(ctxFromTest(t), (expectedStats.Max+expectedStats.Min)/2-1)
+		require.NoError(t, err)
+		err = ctrl.WriteStats(ctxFromTest(t), &stats)
+		require.NoError(t, err)
+		require.Equal(t, (expectedStats.Max+expectedStats.Min)/2, stats.Cap, "failed to Grow")
+
+		err = ctrl.Shrink(ctxFromTest(t), expectedStats.Min)
+		require.NoError(t, err)
+		err = ctrl.WriteStats(ctxFromTest(t), &stats)
+		require.NoError(t, err)
+		require.Equal(t, expectedStats.Min+1, stats.Cap, "failed to Shrink")
+
+		err = ctrl.Clear(ctxFromTest(t))
+		require.NoError(t, err)
+		err = ctrl.WriteStats(ctxFromTest(t), &stats)
+		require.NoError(t, err)
+		require.Equal(t, expectedStats.Min, stats.Cap, "failed to Clear")
+	})
+
+	t.Run("use of send and recv channels with a closed controller", func(t *testing.T) {
+		t.Parallel()
+		send, recv, ctrl := AdaptiveChan[int]()
+
+		ctrl.Close()
+		assertCtrlWriteErr(t, ctrl, ctxFromTest(t), ErrAdaptiveChanControllerClosed)
+
+		sendNonBlock(t, send, ints(10)...)
+		recvNonBlock(t, recv, ints(10)...)
+		close(send)
+		shouldBeClosed(t, recv)
+	})
+}
+
+func TestSendOrErr(t *testing.T) {
+	t.Parallel()
+	const val = 44203
+	var c chan int
+
+	err := sendOrErr(ctxFromTest(t), c, val)
+	require.Error(t, err)
+	require.ErrorIs(t, err, ErrAdaptiveChanControllerClosed)
+
+	c = make(chan int, 1)
+	err = sendOrErr(ctxFromTest(t), c, val)
+	require.NoError(t, err)
+
+	canceledCtx, cancel := context.WithCancel(context.Background())
+	cancel()
+
+	err = sendOrErr(canceledCtx, c, val)
+	require.Error(t, err)
+	require.ErrorIs(t, err, context.Canceled)
+
+	select {
+	case v, ok := <-c:
+		require.True(t, ok)
+		require.Equal(t, val, v)
+	default:
+		t.Fatalf("value not sent to channel")
+	}
+}
+
+func TestRecvOrErr(t *testing.T) {
+	t.Parallel()
+	const (
+		val     = 44203
+		witness = -1
+	)
+	var c chan int
+
+	canceledCtx, cancel := context.WithCancel(context.Background())
+	cancel()
+
+	got := witness
+	err := recvOrErr(canceledCtx, c, &got)
+	require.Error(t, err)
+	require.ErrorIs(t, err, context.Canceled)
+	require.Equal(t, witness, got)
+
+	c = make(chan int, 1)
+	c <- val
+
+	err = recvOrErr(ctxFromTest(t), c, &got)
+	require.NoError(t, err)
+	require.Equal(t, val, got)
+
+	close(c)
+	got = witness
+	err = recvOrErr(ctxFromTest(t), c, &got)
+	require.ErrorIs(t, err, ErrAdaptiveChanClosed)
+	require.Equal(t, witness, got)
+}
+
+// cleanupAC closes the send channel and the controller, and perform a series of
+// rutinary assertions.
+func cleanupAC[T any](t *testing.T, send chan<- T, recv <-chan T, ctrl *AdaptiveChanController) {
+	t.Cleanup(func() {
+		close(send)
+		shouldBeClosed(t, recv)
+
+		var stats AdaptiveChanStats
+		err := ctrl.WriteStats(ctxFromTest(t), &stats)
+		require.Error(t, err)
+		require.ErrorIs(t, err, ErrAdaptiveChanClosed)
+		require.Equal(t, AdaptiveChanStats{}, stats)
+
+		canceledCtx, cancel := context.WithCancel(context.Background())
+		cancel()
+		assertCtrlWriteErr(t, ctrl, canceledCtx, context.Canceled)
+
+		ctrl.Close()
+
+		err = ctrl.WriteStats(ctxFromTest(t), &stats)
+		require.Error(t, err)
+		require.ErrorIs(t, err, ErrAdaptiveChanControllerClosed)
+		require.Equal(t, AdaptiveChanStats{}, stats)
+
+		assertCtrlWriteErr(t, ctrl, ctxFromTest(t), ErrAdaptiveChanControllerClosed)
+	})
+}
+
+func assertCtrlWriteErr(t *testing.T, ctrl *AdaptiveChanController, ctx context.Context, expectedErr error) {
+	t.Helper()
+
+	err := ctrl.Min(ctx, 1)
+	require.Error(t, err)
+	require.ErrorIs(t, err, expectedErr)
+
+	err = ctrl.Max(ctx, 1)
+	require.Error(t, err)
+	require.ErrorIs(t, err, expectedErr)
+
+	err = ctrl.Grow(ctx, 1)
+	require.Error(t, err)
+	require.ErrorIs(t, err, expectedErr)
+
+	err = ctrl.Shrink(ctx, 1)
+	require.Error(t, err)
+	require.ErrorIs(t, err, expectedErr)
+
+	err = ctrl.Clear(ctx)
+	require.Error(t, err)
+	require.ErrorIs(t, err, expectedErr)
+}
+
+func shouldBeClosed[T any](t *testing.T, recv <-chan T) {
+	t.Helper()
+	select {
+	case v, ok := <-recv:
+		require.False(t, ok, "unexpected value %q received", v)
+	case <-ctxFromTest(t).Done():
+		t.Fatalf("context canceled where recv chan should be closed")
+	}
+}
+
+func sendNonBlock[T any](t *testing.T, send chan<- T, s ...T) {
+	t.Helper()
+	var canceled bool
+	for i, v := range s {
+		require.NotPanics(t, func() {
+			select {
+			case send <- v:
+			case <-ctxFromTest(t).Done():
+				canceled = true
+			}
+		})
+		require.False(t, canceled, "context canceled while sending item %d/%d", i+1, len(s))
+	}
+}
+
+func recvNonBlock[T any](t *testing.T, recv <-chan T, s ...T) {
+	t.Helper()
+	var canceled bool
+	for i := range s {
+		select {
+		case s[i] = <-recv:
+		case <-ctxFromTest(t).Done():
+			canceled = true
+		}
+		require.False(t, canceled, "context canceled while receiving item %d/%d", i+1, len(s))
+	}
+}
+
+func ctxFromTest(t *testing.T) context.Context {
+	return ctxFromTestWithDefault(t, time.Second)
+}
+
+func ctxFromTestWithDefault(t *testing.T, d time.Duration) context.Context {
+	require.Greater(t, d, 0*time.Second)
+	deadline, ok := t.Deadline()
+	if !ok {
+		deadline = time.Now().Add(d)
+	}
+	ctx, cancel := context.WithDeadline(context.Background(), deadline)
+	t.Cleanup(cancel)
+	return ctx
+}
diff --git a/pkg/util/ring/ring.go b/pkg/util/ring/ring.go
new file mode 100644
index 00000000000..b4f672b32de
--- /dev/null
+++ b/pkg/util/ring/ring.go
@@ -0,0 +1,254 @@
+package ring
+
+// Ring is a ring buffer backed by a slice that rearranges itself to grow and
+// shrink as needed. It can also be grown and shrunk manually. It is not safe
+// for concurrent use, and the zero value is ready for use. Dequeued and cleared
+// items are zeroed in the underlying slice to release references and allow
+// garbage collection. Leaving growth and shrinkage of the internal slice apart,
+// which can be directly controlled, all operations are allocation free.
+type Ring[T any] struct {
+	buf       []T
+	stats     RingStats
+	back, len int
+
+	// Min sets the minimum capacity that the Ring can have, and takes effect
+	// in the next write operation. The Ring will naturally tend to shrink
+	// towards Min when free capacity allows it. Setting this value has no
+	// immediate effect, but instead affects future writing operations. Min is
+	// valid only if:
+	//	0 < Min && ( Max <= 0 || Min <= Max )
+	// Note that this allows setting Min but not Max, or setting both as well.
+	Min int
+
+	// Max sets the maximum capacity that the Ring can grow to store new items.
+	// Setting this value has no immediate effect, but instead affects future
+	// writing operations. Max is valid only if:
+	//	0 < Max && Min <= Max
+	// Note that this allows setting Max but not Min, or setting both as well.
+	Max int
+}
+
+// RingStats provides general stats for a Ring.
+type RingStats struct {
+	// Len is the used capacity.
+	Len int
+	// Cap is the current total capacity.
+	Cap int
+	// Grown is the number of times a larger buffer was allocated.
+	Grown uint64
+	// Shrunk is the number of times a smaller buffer was allocated.
+	Shrunk uint64
+	// Allocs is Grown + Shrunk.
+	Allocs uint64
+	// Enqueued is the total number of items entered into the Ring, including
+	// those which caused other items to be dropped.
+	Enqueued uint64
+	// Dequeued is the total number of items removed from the Ring, including
+	// items removed with Dequeue and with Clear.
+	Dequeued uint64
+	// Dropped is the number of items lost due to the Ring being at capacity.
+	Dropped uint64
+}
+
+// Len returns the used capacity.
+func (rq *Ring[T]) Len() int {
+	return rq.len
+}
+
+// Cap returns the current total capacity.
+func (rq *Ring[T]) Cap() int {
+	return len(rq.buf)
+}
+
+// WriteStats writes general stats about this Ring to the given *RingStats, if
+// it's non-nil.
+func (rq *Ring[T]) WriteStats(s *RingStats) {
+	if s == nil {
+		return
+	}
+	rq.stats.Len = rq.len
+	rq.stats.Cap = len(rq.buf)
+	*s = rq.stats
+}
+
+// Clear removes all items from the Ring and returns the number of items
+// removed. If Min is valid and Cap() > Min, it will also shrink the capacity to
+// Min. Stats are not cleared, but instead Dequeued is increased by the number
+// of removed items.
+func (rq *Ring[T]) Clear() int {
+	cleared := rq.len
+	rq.stats.Dequeued += uint64(cleared)
+	shouldMigrate := clearShouldMigrate(len(rq.buf), rq.Min, rq.Max)
+
+	if rq.len > 0 && !shouldMigrate {
+		// if we migrate we don't need to clear items, since moving to the new
+		// slice will just have the old slice garbage collected
+		chunk := min(rq.back+rq.len, len(rq.buf))
+		clear(rq.buf[rq.back:chunk])
+		clear(rq.buf[:rq.len-chunk])
+	}
+	rq.back = 0
+	rq.len = 0
+
+	if shouldMigrate {
+		rq.migrate(rq.Min)
+	}
+
+	return cleared
+}
+
+// Shrink makes sure free capacity is not greater than n, shrinking if
+// necessary. If a new allocation is needed then it will be capped to Min, given
+// than Min is valid.
+func (rq *Ring[T]) Shrink(n int) {
+	if n < 0 || rq.len+n >= len(rq.buf) {
+		return
+	}
+	rq.migrate(n)
+}
+
+// Grow makes sure free capacity is at least n, growing if necessary. If a new
+// allocation is needed then it will be capped to Max, given that Max is valid.
+func (rq *Ring[T]) Grow(n int) {
+	if n < 1 || rq.len+n <= len(rq.buf) {
+		return
+	}
+	rq.migrate(n)
+}
+
+func (rq *Ring[T]) migrate(newFreeCap int) {
+	newCap := rq.len + newFreeCap
+	newCap = fixAllocSize(rq.len, rq.Min, rq.Max, newCap)
+	if newCap == len(rq.buf) {
+		return
+	}
+
+	var s []T
+	if newCap > 0 {
+		// if newCap == 0 then just set rq.s to nil
+		s = make([]T, newCap)
+	}
+
+	if len(s) > len(rq.buf) {
+		rq.stats.Grown++
+	} else {
+		rq.stats.Shrunk++
+	}
+
+	if rq.len > 0 {
+		chunk1 := min(rq.back+rq.len, len(rq.buf))
+		copied := copy(s, rq.buf[rq.back:chunk1])
+
+		if copied < rq.len {
+			// wrapped the slice
+			chunk2 := rq.len - copied
+			copy(s[copied:], rq.buf[:chunk2])
+		}
+	}
+
+	rq.back = 0
+	rq.buf = s
+}
+
+// Enqueue adds the given item to the Ring, growing the capacity if needed. If
+// the Ring is at capacity (0 < Max && Min <= Max && rq.Len() == rq.Cap()),
+// then the new item will overwrite the oldest enqueued item.
+func (rq *Ring[T]) Enqueue(v T) {
+	// try to add space if we're at capacity or fix min allocation
+	if rq.len == len(rq.buf) || (minIsValid(rq.Min, rq.Max) && len(rq.buf) < rq.Min) {
+		newFreeCap := rq.len + 1
+		newFreeCap = newFreeCap*3/2 + 1 // classic append: https://go.dev/blog/slices
+		newFreeCap -= rq.len            // migrate only takes free capacity
+		rq.migrate(newFreeCap)
+
+		// if growing was capped at max, then overwrite the first item to be
+		// dequeued
+		if rq.len == len(rq.buf) {
+			rq.stats.Dropped++
+			rq.len--
+			if rq.back++; rq.back >= len(rq.buf) {
+				rq.back = 0 // wrap the slice
+			}
+		}
+	}
+
+	writePos := rq.back + rq.len
+	if writePos >= len(rq.buf) {
+		writePos -= len(rq.buf)
+	}
+
+	rq.buf[writePos] = v
+	rq.len++
+	rq.stats.Enqueued++
+}
+
+// Peek is like Dequeue, but it doesn't remove the item.
+func (rq *Ring[T]) Peek() (v T) {
+	if rq.len == 0 {
+		return
+	}
+
+	return rq.buf[rq.back]
+}
+
+// Dequeue removes the oldest enqueued item and returns it. If the Ring is
+// empty, it returns the zero value.
+func (rq *Ring[T]) Dequeue() (v T) {
+	if rq.len == 0 {
+		return
+	}
+
+	// get the value into v, and also zero out the slice item to release
+	// references so they can be gc'd
+	v, rq.buf[rq.back] = rq.buf[rq.back], v
+	rq.len--
+	if rq.back++; rq.back >= len(rq.buf) {
+		rq.back = 0 // wrap the slice
+	}
+
+	if minIsValid(rq.Min, rq.Max) && rq.len < len(rq.buf)/2+1 {
+		newFreeCap := len(rq.buf)*2/3 + 1 // opposite of growing arithmetic
+		newFreeCap -= rq.len              // migrate only takes free capacity
+		rq.migrate(newFreeCap)
+	}
+	rq.stats.Dequeued++
+
+	return v
+}
+
+// the following functions provide small checks and arithmetics that are far
+// easier to test separately than creating big and more complex tests covering a
+// huge amount of combinatory options. This reduces the complexity of higher
+// level tests and leaves only higher level logic, but also allows us to provide
+// high coverage for even the most rare edge and boundary cases by adding a new
+// line to the test cases table. They're also inlineable, so no penalty in
+// calling them.
+
+func minIsValid(Min, Max int) bool {
+	return 0 < Min && (Max <= 0 || Min <= Max)
+}
+
+func maxIsValid(Min, Max int) bool {
+	return 0 < Max && Min <= Max
+}
+
+func clearShouldMigrate(CurCap, Min, Max int) bool {
+	return minIsValid(Min, Max) && CurCap > Min
+}
+
+// fixAllocSize is a helper to determine what should be the new size to be
+// allocated for a new slice, given the intended NewCap and the current relevant
+// state of Ring. This is expected to be called inside (*Ring).migrate.
+func fixAllocSize(CurLen, Min, Max, NewCap int) int {
+	if minIsValid(Min, Max) { // Min is valid
+		NewCap = max(NewCap, CurLen, Min)
+	} else {
+		NewCap = max(CurLen, NewCap)
+	}
+
+	if maxIsValid(Min, Max) { // Max is valid
+		NewCap = min(NewCap, Max)
+	}
+
+	return NewCap
+}
diff --git a/pkg/util/ring/ring_test.go b/pkg/util/ring/ring_test.go
new file mode 100644
index 00000000000..11f55366937
--- /dev/null
+++ b/pkg/util/ring/ring_test.go
@@ -0,0 +1,573 @@
+package ring
+
+import (
+	"slices"
+	"testing"
+
+	"github.com/stretchr/testify/require"
+)
+
+func ints(n int) []int {
+	ret := make([]int, n)
+	for i := range ret {
+		ret[i] = i + 1
+	}
+	return ret
+}
+
+func TestRing(t *testing.T) {
+	t.Parallel()
+	const (
+		dLen     = 10
+		dHalfLen = dLen / 2
+	)
+	data := ints(dLen)
+	lData := slices.Clone(data[:dHalfLen])
+	rData := slices.Clone(data[dHalfLen:])
+
+	require.NotPanics(t, func() {
+		new(Ring[int]).WriteStats(nil)
+	}, "WriteStats should be panic free")
+
+	t.Run("basic enqueue and dequeue - no min, no max", func(t *testing.T) {
+		t.Parallel()
+		q, expected := new(Ring[int]), new(Ring[int])
+
+		enq(t, q, data...)
+		expected.len = dLen
+		expected.stats.Enqueued = dLen
+		expected.buf = data
+		ringEq(t, expected, q)
+
+		deq(t, q, lData...)
+		expected.back = dHalfLen
+		expected.len = dHalfLen
+		expected.stats.Dequeued = dHalfLen
+		expected.buf = append(make([]int, dHalfLen), rData...)
+		ringEq(t, expected, q)
+
+		enq(t, q, data...)
+		expected.back = 0
+		expected.len = dLen + dHalfLen
+		expected.stats.Enqueued += dLen
+		expected.buf = append(rData, data...)
+		ringEq(t, expected, q)
+
+		deqAll(t, q, append(rData, data...)...)
+		expected.len = 0
+		expected.stats.Dequeued += dLen + dHalfLen
+		expected.buf = []int{}
+		ringEq(t, expected, q)
+
+		enq(t, q, data...)
+		expected.len = dLen
+		expected.stats.Enqueued += dLen
+		expected.buf = data
+		ringEq(t, expected, q)
+
+		clearRing(t, q)
+		expected.len = 0
+		expected.stats.Dequeued += dLen
+		expected.buf = []int{}
+		ringEq(t, expected, q)
+	})
+
+	t.Run("enqueue, dequeue, grow and shrink - no min, yes max", func(t *testing.T) {
+		t.Parallel()
+		q, expected := new(Ring[int]), new(Ring[int])
+		q.Max = dLen
+
+		// basic wrap and overwrite
+		enq(t, q, lData...)
+		enq(t, q, data...)
+		enq(t, q, data...)
+		expected.back = dHalfLen
+		expected.buf = append(rData, lData...)
+		expected.len = dLen
+		expected.stats.Enqueued = 2*dLen + dHalfLen
+		expected.stats.Dropped = dLen + dHalfLen
+		ringEq(t, expected, q)
+		require.Equal(t, dLen, q.Cap())
+
+		// can't allocate past max and cannot shrink because we're at capacity
+		q.Grow(3 * dLen)
+		ringEq(t, expected, q)
+		require.Equal(t, dLen, q.Cap())
+
+		q.Shrink(2 * dLen)
+		ringEq(t, expected, q)
+		require.Equal(t, dLen, q.Cap())
+
+		// remove some items and play with extra space
+		deq(t, q, lData...)
+		expected.back = 0
+		expected.buf = rData
+		expected.len -= dHalfLen
+		expected.stats.Dequeued = dHalfLen
+		require.Equal(t, dLen, q.Cap())
+		ringEq(t, expected, q)
+
+		q.Shrink(1)
+		ringEq(t, expected, q)
+		require.Equal(t, dHalfLen+1, q.Cap())
+
+		q.Grow(2)
+		ringEq(t, expected, q)
+		require.Equal(t, dHalfLen+2, q.Cap())
+
+		q.Grow(dLen)
+		ringEq(t, expected, q)
+		require.Equal(t, dLen, q.Cap())
+	})
+
+	t.Run("enqueue, dequeue, grow and shrink - yes min, no max", func(t *testing.T) {
+		t.Parallel()
+		q, expected := new(Ring[int]), new(Ring[int])
+		q.Min = dHalfLen
+
+		// enqueueing one item should allocate Min
+		enq(t, q, 1)
+		expected.buf = []int{1}
+		expected.len = 1
+		expected.stats.Enqueued = 1
+		ringEq(t, expected, q)
+		require.Equal(t, dHalfLen, q.Cap())
+
+		// clearing should not migrate now
+		clearRing(t, q)
+		expected.buf = []int{}
+		expected.len = 0
+		expected.stats.Dequeued = 1
+		ringEq(t, expected, q)
+		require.Equal(t, dHalfLen, q.Cap())
+
+		// enqueue some data
+		enq(t, q, data...)
+		expected.buf = data
+		expected.len = dLen
+		expected.stats.Enqueued += dLen
+		ringEq(t, expected, q)
+		require.GreaterOrEqual(t, q.Cap(), dLen)
+
+		// now clearing should migrate and move to a slice of Min length
+		clearRing(t, q)
+		expected.buf = []int{}
+		expected.len = 0
+		expected.stats.Dequeued += dLen
+		ringEq(t, expected, q)
+		require.Equal(t, dHalfLen, q.Cap())
+
+		// we shouldn't be able to shrink past Min, but it shouldn't allocate a
+		// greater slice either because it's purpose is to reduce allocated
+		// memory if possible
+		q.Min = dLen
+		q.Shrink(dHalfLen)
+		ringEq(t, expected, q)
+		require.Equal(t, dHalfLen, q.Cap())
+
+		// dequeueing shouldn't allocate either, just in case
+		require.Zero(t, q.Dequeue())
+		ringEq(t, expected, q)
+		require.Equal(t, dHalfLen, q.Cap())
+
+		// enqueueing one item allocates again to Min, which is now greater than
+		// before
+		enq(t, q, 1)
+		expected.buf = []int{1}
+		expected.len = 1
+		expected.stats.Enqueued += 1
+		ringEq(t, expected, q)
+		require.Equal(t, dLen, q.Cap())
+
+		// we reduce Min again, then we should be able to shrink as well
+		q.Min = dHalfLen
+		q.Shrink(dHalfLen)
+		ringEq(t, expected, q)
+		require.Equal(t, dHalfLen+1, q.Cap())
+		q.Shrink(1)
+		ringEq(t, expected, q)
+		require.Equal(t, dHalfLen, q.Cap())
+		q.Shrink(0)
+		ringEq(t, expected, q)
+		require.Equal(t, dHalfLen, q.Cap())
+
+		// enqueue a lot and then dequeue all, we should still see Min cap
+		enq(t, q, data...)
+		expected.buf = append(expected.buf, data...)
+		expected.len += dLen
+		expected.stats.Enqueued += dLen
+		ringEq(t, expected, q)
+		require.GreaterOrEqual(t, q.Cap(), dLen+1)
+
+		deqAll(t, q, expected.buf...)
+		expected.buf = []int{}
+		expected.len = 0
+		expected.stats.Dequeued += dLen + 1
+		ringEq(t, expected, q)
+		require.Equal(t, dHalfLen, q.Cap())
+	})
+
+	t.Run("enqueue, dequeue, grow and shrink - yes min, yes max", func(t *testing.T) {
+		t.Parallel()
+		q, expected := new(Ring[int]), new(Ring[int])
+		q.Min, q.Max = dHalfLen, dLen
+
+		// single enqueueing should allocate for Min
+		enq(t, q, 1)
+		expected.buf = []int{1}
+		expected.len = 1
+		expected.stats.Enqueued = 1
+		ringEq(t, expected, q)
+		require.Equal(t, dHalfLen, q.Cap())
+
+		// enqueue a lot until we overwrite the first item
+		enq(t, q, data...)
+		expected.back = 1
+		expected.buf = append(data[dLen-1:], data[:dLen-1]...)
+		expected.len = dLen
+		expected.stats.Enqueued += dLen
+		expected.stats.Dropped = 1
+		ringEq(t, expected, q)
+		require.Equal(t, dLen, q.Cap())
+
+		// clearing should bring us back to Min alloc
+		clearRing(t, q)
+		expected.back = 0
+		expected.buf = expected.buf[:0]
+		expected.len = 0
+		expected.stats.Dequeued += dLen
+		ringEq(t, expected, q)
+		require.Equal(t, dHalfLen, q.Cap())
+	})
+
+	t.Run("growing and shrinking invariants - no min, no max", func(t *testing.T) {
+		t.Parallel()
+		q, expected := new(Ring[int]), new(Ring[int])
+
+		// dummy grow and shrink
+		q.Grow(0)
+		require.Equal(t, 0, q.Cap())
+		ringEq(t, expected, q)
+
+		q.Shrink(0)
+		require.Equal(t, 0, q.Cap())
+		ringEq(t, expected, q)
+
+		// add 3*dLen and leave 2*dLen
+		q.Grow(3 * dLen)
+		expected.buf = []int{}
+		require.Equal(t, 3*dLen, q.Cap())
+		ringEq(t, expected, q)
+
+		q.Shrink(2 * dLen)
+		require.Equal(t, 2*dLen, q.Cap())
+		ringEq(t, expected, q)
+
+		// add dLen items and play with cap
+		enq(t, q, data...)
+		expected.buf = data
+		expected.len = dLen
+		expected.stats.Enqueued = dLen
+		require.Equal(t, 2*dLen, q.Cap())
+		ringEq(t, expected, q)
+
+		q.Grow(2 * dLen)
+		require.GreaterOrEqual(t, q.Cap(), 3*dLen)
+		ringEq(t, expected, q)
+
+		q.Shrink(0)
+		require.Equal(t, dLen, q.Cap())
+		ringEq(t, expected, q)
+
+		// remove all items and shrink to zero
+		deqAll(t, q, data...)
+		expected.buf = []int{}
+		expected.len = 0
+		expected.stats.Dequeued = dLen
+		require.Equal(t, dLen, q.Cap())
+		ringEq(t, expected, q)
+
+		q.Shrink(0)
+		expected.buf = nil
+		require.Equal(t, 0, q.Cap())
+		ringEq(t, expected, q)
+	})
+}
+
+// enq enqueues the given items into the given Ring.
+func enq[T any](t *testing.T, q *Ring[T], s ...T) {
+	t.Helper()
+
+	initLen := q.Len()
+	initCap := q.Cap()
+	for _, v := range s {
+		require.NotPanics(t, func() {
+			q.Enqueue(v)
+		})
+	}
+
+	expectedLen := initLen + len(s)
+	if q.Max > 0 {
+		expectedMax := max(initCap, q.Max)
+		expectedLen = min(expectedLen, expectedMax)
+	}
+	require.Equal(t, expectedLen, q.Len())
+}
+
+// deq dequeues len(expected) items from the given Ring and compares them to
+// expected. Ring should have at least len(expected) items.
+func deq[T any](t *testing.T, q *Ring[T], expected ...T) {
+	t.Helper()
+
+	if q.Cap() == 0 {
+		require.Nil(t, q.buf)       // internal state
+		require.Equal(t, 0, q.back) // internal state
+		return
+	}
+
+	oldLen := q.Len()
+	require.True(t, oldLen >= len(expected))
+	got := make([]T, len(expected))
+	for i := range got {
+		var val T
+		require.NotPanics(t, func() {
+			prePeekLen := q.Len()
+			val = q.Peek()
+			require.Equal(t, prePeekLen, q.Len())
+			got[i] = q.Dequeue()
+		})
+		require.Equal(t, val, got[i])
+	}
+
+	require.Equal(t, expected, got)
+	require.Equal(t, oldLen-len(expected), q.Len())
+}
+
+// clearRing calls Clear on the given Ring and performs a set of assertions that
+// should be satisfied afterwards.
+func clearRing[T any](t *testing.T, q *Ring[T]) {
+	t.Helper()
+
+	var expectedBuf []T
+	if clearShouldMigrate(q.Cap(), q.Min, q.Max) {
+		expectedBuf = make([]T, q.Min)
+	} else {
+		expectedBuf = make([]T, q.Cap())
+	}
+
+	require.NotPanics(t, func() {
+		q.Clear()
+	})
+	require.Equal(t, expectedBuf, q.buf) // internal state
+	require.Equal(t, 0, q.Len())
+	require.Equal(t, 0, q.back) // internal state
+
+	// dequeueing should yield zero values
+	var zero T
+	for i := 0; i < 10; i++ {
+		var val1, val2 T
+		require.NotPanics(t, func() {
+			val1 = q.Peek()
+			val1 = q.Dequeue()
+		})
+		require.Equal(t, zero, val1)
+		require.Equal(t, zero, val2)
+	}
+}
+
+// deqAll depletes the given Ring and compares the dequeued items to those
+// provided.
+func deqAll[T any](t *testing.T, q *Ring[T], expected ...T) {
+	t.Helper()
+
+	deq[T](t, q, expected...)
+
+	zeroS := make([]T, q.Cap())
+	require.Equal(t, zeroS, q.buf) // internal state
+
+	require.Equal(t, 0, q.Len())
+
+	// dequeueing further should yield zero values when empty
+	var zero T
+	for i := 0; i < 10; i++ {
+		var val1, val2 T
+		require.NotPanics(t, func() {
+			val1 = q.Peek()
+			val2 = q.Dequeue()
+		})
+		require.Equal(t, zero, val1)
+		require.Equal(t, zero, val2)
+	}
+
+	clearRing(t, q)
+}
+
+// ringEq tests that the given Rings are the same in many aspects. The following
+// are the things that are not checked:
+//   - The values of Min and Max, since the code does not programmatically
+//     channge them
+//   - Allocation numbers (Cap, Grown, Shrunk, Allocs)
+//   - The free capacity to the right of `got`
+func ringEq[T any](t *testing.T, expected, got *Ring[T]) {
+	t.Helper()
+
+	var expStats, gotStats RingStats
+	require.NotPanics(t, func() {
+		expected.WriteStats(&expStats)
+		got.WriteStats(&gotStats)
+	})
+
+	// capacity and allocations are to be tested separately
+	removeAllocStats(&expStats)
+	removeAllocStats(&gotStats)
+
+	require.Equal(t, expStats, gotStats, "expStats == gotStats")
+
+	// internal state
+	require.Equal(t, expected.back, got.back, "expected.back == got.back")
+	// only check for used capacity
+	require.Equal(t, expected.buf, got.buf[:min(got.back+got.len, len(got.buf))],
+		"expected.buf == got.buf[:min(got.back+got.len, len(got.s))]")
+}
+
+func removeAllocStats(s *RingStats) {
+	s.Cap = 0
+	s.Grown = 0
+	s.Shrunk = 0
+	s.Allocs = 0
+}
+
+func TestMinMaxValidity(t *testing.T) {
+	t.Parallel()
+
+	testCases := []struct {
+		Min, Max               int
+		minIsValid, maxIsValid bool
+	}{
+		{Min: 0, Max: 0, minIsValid: false, maxIsValid: false},
+		{Min: 0, Max: 1, minIsValid: false, maxIsValid: true},
+		{Min: 0, Max: 2, minIsValid: false, maxIsValid: true},
+		{Min: 1, Max: 0, minIsValid: true, maxIsValid: false},
+		{Min: 1, Max: 1, minIsValid: true, maxIsValid: true},
+		{Min: 1, Max: 2, minIsValid: true, maxIsValid: true},
+		{Min: 2, Max: 0, minIsValid: true, maxIsValid: false},
+		{Min: 2, Max: 1, minIsValid: false, maxIsValid: false},
+		{Min: 2, Max: 2, minIsValid: true, maxIsValid: true},
+	}
+
+	for i, tc := range testCases {
+		gotMinIsValid := minIsValid(tc.Min, tc.Max)
+		require.Equal(t, tc.minIsValid, gotMinIsValid,
+			"test index %d; test data: %#v", i, tc)
+
+		gotMaxIsValid := maxIsValid(tc.Min, tc.Max)
+		require.Equal(t, tc.maxIsValid, gotMaxIsValid,
+			"test index %d; test data: %#v", i, tc)
+	}
+}
+
+func TestClearShouldMigrate(t *testing.T) {
+	t.Parallel()
+
+	testCases := []struct {
+		// we don't need to include Max in the test, we just disable it by
+		// passing zero because Max is only needed to establish the validity of
+		// Min. The validity of Min wrt Max is already covered in the test for
+		// minIsValid, and once Min is valid Max has no impact on the outcome of
+		// clearShouldMigrate.
+		CurCap, Min int
+		expected    bool
+	}{
+		{CurCap: 0, Min: 0, expected: false},
+		{CurCap: 0, Min: 9, expected: false},
+		{CurCap: 0, Min: 10, expected: false},
+		{CurCap: 0, Min: 11, expected: false},
+		{CurCap: 10, Min: 0, expected: false},
+		{CurCap: 10, Min: 9, expected: true},
+		{CurCap: 10, Min: 10, expected: false},
+		{CurCap: 10, Min: 11, expected: false},
+	}
+
+	for i, tc := range testCases {
+		got := clearShouldMigrate(tc.CurCap, tc.Min, 0)
+		require.Equal(t, tc.expected, got,
+			"test index %d; test data: %#v", i, tc)
+	}
+}
+
+func TestFixAllocSize(t *testing.T) {
+	t.Parallel()
+
+	testCases := []struct {
+		CurLen, Min, Max, NewCap, expected int
+	}{
+		// we don't need to add test cases for odd configurations of Min and Max
+		// not being valid for different reasons because that is already covered
+		// in the unit tests for minIsValid and maxIsValid. It suffices to
+		// provide a zero for Min or Max to disable their respective behaviour
+
+		{CurLen: 0, Min: 0, Max: 0, NewCap: 0, expected: 0},
+		{CurLen: 0, Min: 0, Max: 0, NewCap: 5, expected: 5},
+
+		{CurLen: 0, Min: 0, Max: 10, NewCap: 0, expected: 0},
+		{CurLen: 0, Min: 0, Max: 10, NewCap: 9, expected: 9},
+		{CurLen: 0, Min: 0, Max: 10, NewCap: 10, expected: 10},
+		{CurLen: 0, Min: 0, Max: 10, NewCap: 11, expected: 10},
+
+		{CurLen: 0, Min: 10, Max: 0, NewCap: 0, expected: 10},
+		{CurLen: 0, Min: 10, Max: 0, NewCap: 5, expected: 10},
+		{CurLen: 0, Min: 10, Max: 0, NewCap: 9, expected: 10},
+		{CurLen: 0, Min: 10, Max: 0, NewCap: 10, expected: 10},
+		{CurLen: 0, Min: 10, Max: 0, NewCap: 11, expected: 11},
+
+		{CurLen: 0, Min: 10, Max: 10, NewCap: 0, expected: 10},
+		{CurLen: 0, Min: 10, Max: 10, NewCap: 5, expected: 10},
+		{CurLen: 0, Min: 10, Max: 10, NewCap: 9, expected: 10},
+		{CurLen: 0, Min: 10, Max: 10, NewCap: 10, expected: 10},
+		{CurLen: 0, Min: 10, Max: 10, NewCap: 11, expected: 10},
+
+		{CurLen: 0, Min: 10, Max: 20, NewCap: 0, expected: 10},
+		{CurLen: 0, Min: 10, Max: 20, NewCap: 5, expected: 10},
+		{CurLen: 0, Min: 10, Max: 20, NewCap: 9, expected: 10},
+		{CurLen: 0, Min: 10, Max: 20, NewCap: 10, expected: 10},
+		{CurLen: 0, Min: 10, Max: 20, NewCap: 19, expected: 19},
+		{CurLen: 0, Min: 10, Max: 20, NewCap: 20, expected: 20},
+		{CurLen: 0, Min: 10, Max: 20, NewCap: 21, expected: 20},
+
+		{CurLen: 5, Min: 0, Max: 0, NewCap: 0, expected: 5},
+		{CurLen: 5, Min: 0, Max: 0, NewCap: 5, expected: 5},
+		{CurLen: 5, Min: 0, Max: 0, NewCap: 10, expected: 10},
+
+		{CurLen: 5, Min: 0, Max: 10, NewCap: 0, expected: 5},
+		{CurLen: 5, Min: 0, Max: 10, NewCap: 5, expected: 5},
+		{CurLen: 5, Min: 0, Max: 10, NewCap: 9, expected: 9},
+		{CurLen: 5, Min: 0, Max: 10, NewCap: 10, expected: 10},
+		{CurLen: 5, Min: 0, Max: 10, NewCap: 11, expected: 10},
+
+		{CurLen: 5, Min: 10, Max: 0, NewCap: 0, expected: 10},
+		{CurLen: 5, Min: 10, Max: 0, NewCap: 5, expected: 10},
+		{CurLen: 5, Min: 10, Max: 0, NewCap: 9, expected: 10},
+		{CurLen: 5, Min: 10, Max: 0, NewCap: 10, expected: 10},
+		{CurLen: 5, Min: 10, Max: 0, NewCap: 11, expected: 11},
+
+		{CurLen: 5, Min: 10, Max: 10, NewCap: 0, expected: 10},
+		{CurLen: 5, Min: 10, Max: 10, NewCap: 5, expected: 10},
+		{CurLen: 5, Min: 10, Max: 10, NewCap: 9, expected: 10},
+		{CurLen: 5, Min: 10, Max: 10, NewCap: 10, expected: 10},
+		{CurLen: 5, Min: 10, Max: 10, NewCap: 11, expected: 10},
+
+		{CurLen: 5, Min: 10, Max: 20, NewCap: 0, expected: 10},
+		{CurLen: 5, Min: 10, Max: 20, NewCap: 5, expected: 10},
+		{CurLen: 5, Min: 10, Max: 20, NewCap: 9, expected: 10},
+		{CurLen: 5, Min: 10, Max: 20, NewCap: 10, expected: 10},
+		{CurLen: 5, Min: 10, Max: 20, NewCap: 19, expected: 19},
+		{CurLen: 5, Min: 10, Max: 20, NewCap: 20, expected: 20},
+		{CurLen: 5, Min: 10, Max: 20, NewCap: 21, expected: 20},
+	}
+
+	for i, tc := range testCases {
+		got := fixAllocSize(tc.CurLen, tc.Min, tc.Max, tc.NewCap)
+		require.Equal(t, tc.expected, got,
+			"test index %d; test data %#v", i, tc)
+	}
+}