diff --git a/sync3/caches/user.go b/sync3/caches/user.go
index 51f18a22..3a7ec0c1 100644
--- a/sync3/caches/user.go
+++ b/sync3/caches/user.go
@@ -428,7 +428,9 @@ func (c *UserCache) Invites() map[string]UserRoomData {
 }
 
 // AttemptToFetchPrevBatch tries to find a prev_batch value for the given event. This may not always succeed.
-func (c *UserCache) AttemptToFetchPrevBatch(roomID string, firstTimelineEvent *EventData) (prevBatch string) {
+func (c *UserCache) AttemptToFetchPrevBatch(ctx context.Context, roomID string, firstTimelineEvent *EventData) (prevBatch string) {
+	_, span := internal.StartSpan(ctx, "AttemptToFetchPrevBatch")
+	defer span.End()
 	return c.store.GetClosestPrevBatch(roomID, firstTimelineEvent.NID)
 }
 
diff --git a/sync3/conn.go b/sync3/conn.go
index 5b8e95c5..7fcb32e0 100644
--- a/sync3/conn.go
+++ b/sync3/conn.go
@@ -135,6 +135,7 @@ func (c *Conn) isOutstanding(pos int64) bool {
 // client. It will NOT be reported to Sentry---this should happen as close as possible
 // to the creation of the error (or else Sentry cannot provide a meaningful traceback.)
 func (c *Conn) OnIncomingRequest(ctx context.Context, req *Request, start time.Time) (resp *Response, herr *internal.HandlerError) {
+	ctx, span := internal.StartSpan(ctx, "OnIncomingRequest.AcquireMutex")
 	c.cancelOutstandingRequestMu.Lock()
 	if c.cancelOutstandingRequest != nil {
 		c.cancelOutstandingRequest()
@@ -149,6 +150,7 @@ func (c *Conn) OnIncomingRequest(ctx context.Context, req *Request, start time.T
 	// it's intentional for the lock to be held whilst inside HandleIncomingRequest
 	// as it guarantees linearisation of data within a single connection
 	defer c.mu.Unlock()
+	span.End()
 
 	isFirstRequest := req.pos == 0
 	isRetransmit := !isFirstRequest && c.lastClientRequest.pos == req.pos
diff --git a/sync3/extensions/extensions.go b/sync3/extensions/extensions.go
index be0cdc92..4eebd4c9 100644
--- a/sync3/extensions/extensions.go
+++ b/sync3/extensions/extensions.go
@@ -318,7 +318,9 @@ func (h *Handler) HandleLiveUpdate(ctx context.Context, update caches.Update, re
 	extCtx.Handler = h
 	exts := req.EnabledExtensions()
 	for _, ext := range exts {
-		ext.AppendLive(ctx, res, extCtx, update)
+		childCtx, region := internal.StartSpan(ctx, "extension_live_"+ext.Name())
+		ext.AppendLive(childCtx, res, extCtx, update)
+		region.End()
 	}
 }
 
diff --git a/sync3/handler/connstate_live.go b/sync3/handler/connstate_live.go
index ef2a0da8..70da78db 100644
--- a/sync3/handler/connstate_live.go
+++ b/sync3/handler/connstate_live.go
@@ -61,6 +61,7 @@ func (s *connStateLive) liveUpdate(
 	// block until we get a new event, with appropriate timeout
 	startTime := time.Now()
 	hasLiveStreamed := false
+	numProcessedUpdates := 0
 	for response.ListOps() == 0 && len(response.Rooms) == 0 && !response.Extensions.HasData(isInitial) {
 		hasLiveStreamed = true
 		timeToWait := time.Duration(req.TimeoutMSecs()) * time.Millisecond
@@ -82,10 +83,12 @@ func (s *connStateLive) liveUpdate(
 			return
 		case update := <-s.updates:
 			s.processUpdate(ctx, update, response, ex)
+			numProcessedUpdates++
 			// if there's more updates and we don't have lots stacked up already, go ahead and process another
-			for len(s.updates) > 0 && response.ListOps() < 50 {
+			for len(s.updates) > 0 && numProcessedUpdates < 100 {
 				update = <-s.updates
 				s.processUpdate(ctx, update, response, ex)
+				numProcessedUpdates++
 			}
 		}
 	}
@@ -128,6 +131,8 @@ func (s *connStateLive) processUpdate(ctx context.Context, update caches.Update,
 }
 
 func (s *connStateLive) processLiveUpdate(ctx context.Context, up caches.Update, response *sync3.Response) bool {
+	_, span := internal.StartSpan(ctx, "processLiveUpdate")
+	defer span.End()
 	internal.AssertWithContext(ctx, "processLiveUpdate: response list length != internal list length", s.lists.Len() == len(response.Lists))
 	internal.AssertWithContext(ctx, "processLiveUpdate: request list length != internal list length", s.lists.Len() == len(s.muxedReq.Lists))
 	roomUpdate, _ := up.(caches.RoomUpdate)
@@ -236,7 +241,7 @@ func (s *connStateLive) processLiveUpdate(ctx context.Context, up caches.Update,
 				})
 				if len(r.Timeline) == 0 && r.PrevBatch == "" {
 					// attempt to fill in the prev_batch value for this room
-					prevBatch := s.userCache.AttemptToFetchPrevBatch(roomEventUpdate.RoomID(), roomEventUpdate.EventData)
+					prevBatch := s.userCache.AttemptToFetchPrevBatch(ctx, roomEventUpdate.RoomID(), roomEventUpdate.EventData)
 					if prevBatch != "" {
 						r.PrevBatch = prevBatch
 					}
@@ -246,11 +251,13 @@ func (s *connStateLive) processLiveUpdate(ctx context.Context, up caches.Update,
 				sender := roomEventUpdate.EventData.Sender
 				if s.lazyCache.IsLazyLoading(roomID) && !s.lazyCache.IsSet(roomID, sender) {
 					// load the state event
+					_, span := internal.StartSpan(ctx, "LazyLoadingMemberEvent")
 					memberEvent := s.globalCache.LoadStateEvent(context.Background(), roomID, s.loadPositions[roomID], "m.room.member", sender)
 					if memberEvent != nil {
 						r.RequiredState = append(r.RequiredState, memberEvent)
 						s.lazyCache.AddUser(roomID, sender)
 					}
+					span.End()
 				}
 			}
 		}