diff --git a/sentry_sdk/_span_batcher.py b/sentry_sdk/_span_batcher.py
index 2033cf7845..deadf1b40e 100644
--- a/sentry_sdk/_span_batcher.py
+++ b/sentry_sdk/_span_batcher.py
@@ -21,10 +21,10 @@ class SpanBatcher(Batcher["SpanJSON"]):
     # MAX_BEFORE_FLUSH should be lower than MAX_BEFORE_DROP, so that there is
     # a bit of a buffer for spans that appear between the trigger to flush
     # and actually flushing the buffer.
-    #
-    # The max limits are all per trace (per bucket).
     MAX_ENVELOPE_SIZE = 1000  # spans
+
     MAX_BEFORE_FLUSH = 1000
+    GLOBAL_MAX_BEFORE_FLUSH = 5_000
 
     MAX_BEFORE_DROP = 2000
     GLOBAL_MAX_BEFORE_DROP = 10_000
@@ -105,9 +105,13 @@ def _flush_loop(self) -> None:
 
             self._flush(only_pending=True)
 
-            if self._total_running_size >= self.GLOBAL_MAX_BYTES_BEFORE_FLUSH or (
-                time.monotonic() - self._last_full_flush
-                >= self.FLUSH_WAIT_TIME + jitter
+            if (
+                self._span_number >= self.GLOBAL_MAX_BEFORE_FLUSH
+                or self._total_running_size >= self.GLOBAL_MAX_BYTES_BEFORE_FLUSH
+                or (
+                    time.monotonic() - self._last_full_flush
+                    >= self.FLUSH_WAIT_TIME + jitter
+                )
             ):
                 self._flush()
                 self._last_full_flush = time.monotonic()
@@ -155,7 +159,9 @@ def add(self, span: "SpanJSON") -> None:
                     notify = True
                 else:
                     notify = (
-                        self._total_running_size >= self.GLOBAL_MAX_BYTES_BEFORE_FLUSH
+                        self._span_number >= self.GLOBAL_MAX_BEFORE_FLUSH
+                        or self._total_running_size
+                        >= self.GLOBAL_MAX_BYTES_BEFORE_FLUSH
                     )
 
             if notify:
diff --git a/tests/tracing/test_span_batcher.py b/tests/tracing/test_span_batcher.py
index fa34ab26e1..172117e742 100644
--- a/tests/tracing/test_span_batcher.py
+++ b/tests/tracing/test_span_batcher.py
@@ -339,6 +339,67 @@ def test_weight_based_flushing_by_attribute_size(
 
 
 def test_global_length_based_flushing(sentry_init, capture_items, monkeypatch):
+    """A flush event is triggered when the batcher contains GLOBAL_MAX_BEFORE_FLUSH spans."""
+    monkeypatch.setattr(SpanBatcher, "GLOBAL_MAX_BEFORE_FLUSH", 2)
+    # set the time-based flush limit to something huge so that we're not hitting
+    # it since we want to test GLOBAL_MAX_BEFORE_FLUSH instead
+    monkeypatch.setattr(SpanBatcher, "FLUSH_WAIT_TIME", 100000)
+
+    sentry_init(
+        traces_sample_rate=1.0,
+        trace_lifecycle="stream",
+    )
+
+    items = capture_items("span")
+
+    with sentry_sdk.traces.start_span(name="span"):
+        pass
+
+    sentry_sdk.traces.new_trace()
+    with sentry_sdk.traces.start_span(name="span 2"):
+        pass
+
+    time.sleep(0.1)
+
+    assert len(items) == 2
+    assert items[0].payload["name"] == "span"
+
+
+def test_span_number_reset_after_length_based_flushing(
+    sentry_init, capture_items, monkeypatch
+):
+    """Span is not flushed after a flush reduces the number of spans in the batcher below the global limit."""
+    monkeypatch.setattr(SpanBatcher, "GLOBAL_MAX_BEFORE_FLUSH", 2)
+    # set the time-based flush limit to something huge so that we're not hitting
+    # it since we want to test GLOBAL_MAX_BYTES_BEFORE_FLUSH instead
+    monkeypatch.setattr(SpanBatcher, "FLUSH_WAIT_TIME", 100000)
+
+    sentry_init(
+        traces_sample_rate=1.0,
+        trace_lifecycle="stream",
+    )
+
+    items = capture_items("span")
+
+    with sentry_sdk.traces.start_span(name="span"):
+        pass
+
+    sentry_sdk.traces.new_trace()
+    with sentry_sdk.traces.start_span(name="span"):
+        pass
+
+    time.sleep(0.1)
+
+    with sentry_sdk.traces.start_span(name="span"):
+        pass
+
+    time.sleep(0.1)
+
+    assert len(items) == 2
+    assert items[0].payload["name"] == "span"
+
+
+def test_global_weight_based_flushing(sentry_init, capture_items, monkeypatch):
     """When the batcher reaches GLOBAL_MAX_BYTES_BEFORE_FLUSH, all buckets will be flushed."""
     # Limit of 2_000 is just above the size of a bare span.
     monkeypatch.setattr(SpanBatcher, "GLOBAL_MAX_BYTES_BEFORE_FLUSH", 2_000)
@@ -366,7 +427,7 @@ def test_global_length_based_flushing(sentry_init, capture_items, monkeypatch):
     assert items[0].payload["name"] == "span"
 
 
-def test_total_size_reset_after_length_based_flushing(
+def test_total_size_reset_after_weight_based_flushing(
     sentry_init, capture_items, monkeypatch
 ):
     """Span is not flushed after a flush reduces the combined span size in bytes below the global limit."""