feat(cdk): add sentry_sdk.capture_message() to MemoryMonitor for high-memory warnings (AI-Triage PR)

airbyte_cdk/utils/memory_monitor.py

@@ -8,6 +8,11 @@
 from pathlib import Path
 from typing import Optional
 
+try:
+    import sentry_sdk
+except ImportError:
+    sentry_sdk = None  # type: ignore[assignment]
+
 logger = logging.getLogger("airbyte")
 
 # cgroup v2 paths
@@ -47,6 +52,7 @@ def __init__(
         self._message_count = 0
         self._cgroup_version: Optional[int] = None
         self._probed = False
+        self._sentry_alerted = False
 
     def _probe_cgroup(self) -> None:
         """Detect which cgroup version (if any) is available.
@@ -138,3 +144,10 @@ def check_memory_usage(self) -> None:
                 usage_gb,
                 limit_gb,
             )
+            if not self._sentry_alerted and sentry_sdk is not None:
+                self._sentry_alerted = True
+                sentry_sdk.capture_message(
+                    "Source memory usage at %d%% of container limit (%.2f / %.2f GB)."
+                    % (usage_percent, usage_gb, limit_gb),
+                    level="warning",
+                )

pyproject.toml

@@ -235,7 +235,8 @@ exclude = [
 DEP001 = [
   # These are imported but not declared:
   "source_declarative_manifest",  # Imported only in dynamic import tests, not in main code
-  "google"  # Imported via google.cloud.secretmanager_v1 which is provided by google-cloud-secret-manager
+  "google",  # Imported via google.cloud.secretmanager_v1 which is provided by google-cloud-secret-manager
+  "sentry_sdk",  # Soft-imported in memory_monitor.py; provided at runtime by connector-execution-service
 ]
 
 # DEP002: Project should not contain unused dependencies.

unit_tests/utils/test_memory_monitor.py

@@ -4,7 +4,7 @@
 
 import logging
 from pathlib import Path
-from unittest.mock import patch
+from unittest.mock import MagicMock, patch
 
 import pytest
 
@@ -245,3 +245,74 @@ def mock_read_text(self: Path) -> str:
     ):
         monitor.check_memory_usage()
     assert not caplog.records
+
+
+# ---------------------------------------------------------------------------
+# check_memory_usage — Sentry capture_message
+# ---------------------------------------------------------------------------
+
+
+def test_sentry_capture_message_called_on_high_memory() -> None:
+    """sentry_sdk.capture_message() should be called once when memory exceeds 90%."""
+    mock_capture = MagicMock()
+    monitor = MemoryMonitor(check_interval=1)
+    with (
+        patch.object(Path, "exists", _v2_exists),
+        patch.object(Path, "read_text", _v2_mock_read(usage=_MOCK_USAGE_AT_90)),
+        patch("airbyte_cdk.utils.memory_monitor.sentry_sdk") as mock_sentry,
+    ):
+        mock_sentry.capture_message = mock_capture
+        monitor.check_memory_usage()
+
+    mock_capture.assert_called_once()
+    call_args = mock_capture.call_args
+    assert "91%" in call_args[0][0]
+    assert call_args[1]["level"] == "warning"
+
+
+def test_sentry_capture_message_only_once_per_sync() -> None:
+    """sentry_sdk.capture_message() should fire only once even if memory stays high."""
+    mock_capture = MagicMock()
+    monitor = MemoryMonitor(check_interval=1)
+    with (
+        patch.object(Path, "exists", _v2_exists),
+        patch.object(Path, "read_text", _v2_mock_read(usage=_MOCK_USAGE_AT_90)),
+        patch("airbyte_cdk.utils.memory_monitor.sentry_sdk") as mock_sentry,
+    ):
+        mock_sentry.capture_message = mock_capture
+        monitor.check_memory_usage()
+        monitor.check_memory_usage()
+        monitor.check_memory_usage()
+
+    mock_capture.assert_called_once()
+
+
+def test_sentry_not_called_below_threshold() -> None:
+    """sentry_sdk.capture_message() should not be called when memory is below 90%."""
+    mock_capture = MagicMock()
+    monitor = MemoryMonitor(check_interval=1)
+    with (
+        patch.object(Path, "exists", _v2_exists),
+        patch.object(Path, "read_text", _v2_mock_read(usage=_MOCK_USAGE_BELOW)),
+        patch("airbyte_cdk.utils.memory_monitor.sentry_sdk") as mock_sentry,
+    ):
+        mock_sentry.capture_message = mock_capture
+        monitor.check_memory_usage()
+
+    mock_capture.assert_not_called()
+
+
+def test_sentry_unavailable_degrades_gracefully(caplog: pytest.LogCaptureFixture) -> None:
+    """When sentry_sdk is None (not installed), warning log should still be emitted."""
+    monitor = MemoryMonitor(check_interval=1)
+    with (
+        caplog.at_level(logging.WARNING, logger="airbyte"),
+        patch.object(Path, "exists", _v2_exists),
+        patch.object(Path, "read_text", _v2_mock_read(usage=_MOCK_USAGE_AT_90)),
+        patch("airbyte_cdk.utils.memory_monitor.sentry_sdk", None),
+    ):
+        monitor.check_memory_usage()
+
+    # Warning log should still be emitted even when sentry_sdk is unavailable
+    assert len(caplog.records) == 1
+    assert "91%" in caplog.records[0].message