feat(pymongo): add PyMongo integration (#1590)

* feat(pymongo): add PyMongo integration

Adds breadcrumbs and performance traces for PyMongo queries using an
official monitoring API. Integration is similar to the one available in
OpenTelemetry, tags set to values recommended for attributes by OT as
specified in `Span Operations` guidelines.

Personal identifiable information (PII) will be stripped from all PyMongo commands. (This was tested in the PyMongo versions below, but "should" also be future proof)

PyMongo version selection explanation:
* 3.1 - introduction of monitoring API. Only Python 2.7 and 3.6
supported.
* 3.12 - latest 3.x release, support for 2.7, 3.6-3.9 (3.7-3.9 added in
various minor releases between 3.1 and 3.12).
* 4.0 - no support for 2.7, added support for 3.10.
* 4.1 - no support for 3.6.0-3.6.1.
* 4.2 - no support for any 3.6.

Co-authored-by: Szymon Soloch <ssoloch@opera.com>
Co-authored-by: Anton Pirker <anton.pirker@sentry.io>

.github/workflows/test-integration-pymongo.yml

@@ -0,0 +1,62 @@
+name: Test pymongo
+
+on:
+  push:
+    branches:
+      - master
+      - release/**
+
+  pull_request:
+
+# Cancel in progress workflows on pull_requests.
+# https://docs.github.com/en/actions/using-jobs/using-concurrency#example-using-a-fallback-value
+concurrency:
+  group: ${{ github.workflow }}-${{ github.head_ref || github.run_id }}
+  cancel-in-progress: true
+
+permissions:
+  contents: read
+
+env:
+  BUILD_CACHE_KEY: ${{ github.sha }}
+  CACHED_BUILD_PATHS: |
+    ${{ github.workspace }}/dist-serverless
+
+jobs:
+  test:
+    name: pymongo, python ${{ matrix.python-version }}, ${{ matrix.os }}
+    runs-on: ${{ matrix.os }}
+    timeout-minutes: 45
+    continue-on-error: true
+
+    strategy:
+      matrix:
+        python-version: ["2.7","3.6","3.7","3.8","3.9","3.10"]
+        os: [ubuntu-latest]
+
+    steps:
+      - uses: actions/checkout@v3
+      - uses: actions/setup-python@v4
+        with:
+          python-version: ${{ matrix.python-version }}
+
+      - name: Setup Test Env
+        env:
+          PGHOST: localhost
+          PGPASSWORD: sentry
+        run: |
+          pip install codecov tox
+
+      - name: Test pymongo
+        env:
+          CI_PYTHON_VERSION: ${{ matrix.python-version }}
+        timeout-minutes: 45
+        shell: bash
+        run: |
+          set -x # print commands that are executed
+          coverage erase
+
+          ./scripts/runtox.sh "${{ matrix.python-version }}-pymongo" --cov=tests --cov=sentry_sdk --cov-report= --cov-branch
+          coverage combine .coverage*
+          coverage xml -i
+          codecov --file coverage.xml

linter-requirements.txt

@@ -4,6 +4,7 @@ flake8==5.0.4
 types-certifi
 types-redis
 types-setuptools
+pymongo # There is no separate types module.
 flake8-bugbear==22.9.23
 pep8-naming==0.13.2
 pre-commit # local linting

sentry_sdk/integrations/pymongo.py

@@ -0,0 +1,183 @@
+from __future__ import absolute_import
+import copy
+
+from sentry_sdk import Hub
+from sentry_sdk.hub import _should_send_default_pii
+from sentry_sdk.integrations import DidNotEnable, Integration
+from sentry_sdk.tracing import Span
+from sentry_sdk.utils import capture_internal_exceptions
+
+from sentry_sdk._types import MYPY
+
+try:
+    from pymongo import monitoring
+except ImportError:
+    raise DidNotEnable("Pymongo not installed")
+
+if MYPY:
+    from typing import Any, Dict, Union
+
+    from pymongo.monitoring import (
+        CommandFailedEvent,
+        CommandStartedEvent,
+        CommandSucceededEvent,
+    )
+
+
+SAFE_COMMAND_ATTRIBUTES = [
+    "insert",
+    "ordered",
+    "find",
+    "limit",
+    "singleBatch",
+    "aggregate",
+    "createIndexes",
+    "indexes",
+    "delete",
+    "findAndModify",
+    "renameCollection",
+    "to",
+    "drop",
+]
+
+
+def _strip_pii(command):
+    # type: (Dict[str, Any]) -> Dict[str, Any]
+    for key in command:
+        is_safe_field = key in SAFE_COMMAND_ATTRIBUTES
+        if is_safe_field:
+            # Skip if safe key
+            continue
+
+        update_db_command = key == "update" and "findAndModify" not in command
+        if update_db_command:
+            # Also skip "update" db command because it is save.
+            # There is also an "update" key in the "findAndModify" command, which is NOT safe!
+            continue
+
+        # Special stripping for documents
+        is_document = key == "documents"
+        if is_document:
+            for doc in command[key]:
+                for doc_key in doc:
+                    doc[doc_key] = "%s"
+            continue
+
+        # Special stripping for dict style fields
+        is_dict_field = key in ["filter", "query", "update"]
+        if is_dict_field:
+            for item_key in command[key]:
+                command[key][item_key] = "%s"
+            continue
+
+        # For pipeline fields strip the `$match` dict
+        is_pipeline_field = key == "pipeline"
+        if is_pipeline_field:
+            for pipeline in command[key]:
+                for match_key in pipeline["$match"] if "$match" in pipeline else []:
+                    pipeline["$match"][match_key] = "%s"
+            continue
+
+        # Default stripping
+        command[key] = "%s"
+
+    return command
+
+
+class CommandTracer(monitoring.CommandListener):
+    def __init__(self):
+        # type: () -> None
+        self._ongoing_operations = {}  # type: Dict[int, Span]
+
+    def _operation_key(self, event):
+        # type: (Union[CommandFailedEvent, CommandStartedEvent, CommandSucceededEvent]) -> int
+        return event.request_id
+
+    def started(self, event):
+        # type: (CommandStartedEvent) -> None
+        hub = Hub.current
+        if hub.get_integration(PyMongoIntegration) is None:
+            return
+        with capture_internal_exceptions():
+            command = dict(copy.deepcopy(event.command))
+
+            command.pop("$db", None)
+            command.pop("$clusterTime", None)
+            command.pop("$signature", None)
+
+            op = "db.query"
+
+            tags = {
+                "db.name": event.database_name,
+                "db.system": "mongodb",
+                "db.operation": event.command_name,
+            }
+
+            try:
+                tags["net.peer.name"] = event.connection_id[0]
+                tags["net.peer.port"] = str(event.connection_id[1])
+            except TypeError:
+                pass
+
+            data = {"operation_ids": {}}  # type: Dict[str, Dict[str, Any]]
+
+            data["operation_ids"]["operation"] = event.operation_id
+            data["operation_ids"]["request"] = event.request_id
+
+            try:
+                lsid = command.pop("lsid")["id"]
+                data["operation_ids"]["session"] = str(lsid)
+            except KeyError:
+                pass
+
+            if not _should_send_default_pii():
+                command = _strip_pii(command)
+
+            query = "{} {}".format(event.command_name, command)
+            span = hub.start_span(op=op, description=query)
+
+            for tag, value in tags.items():
+                span.set_tag(tag, value)
+
+            for key, value in data.items():
+                span.set_data(key, value)
+
+            with capture_internal_exceptions():
+                hub.add_breadcrumb(message=query, category="query", type=op, data=tags)
+
+            self._ongoing_operations[self._operation_key(event)] = span.__enter__()
+
+    def failed(self, event):
+        # type: (CommandFailedEvent) -> None
+        hub = Hub.current
+        if hub.get_integration(PyMongoIntegration) is None:
+            return
+
+        try:
+            span = self._ongoing_operations.pop(self._operation_key(event))
+            span.set_status("internal_error")
+            span.__exit__(None, None, None)
+        except KeyError:
+            return
+
+    def succeeded(self, event):
+        # type: (CommandSucceededEvent) -> None
+        hub = Hub.current
+        if hub.get_integration(PyMongoIntegration) is None:
+            return
+
+        try:
+            span = self._ongoing_operations.pop(self._operation_key(event))
+            span.set_status("ok")
+            span.__exit__(None, None, None)
+        except KeyError:
+            pass
+
+
+class PyMongoIntegration(Integration):
+    identifier = "pymongo"
+
+    @staticmethod
+    def setup_once():
+        # type: () -> None
+        monitoring.register(CommandTracer())

setup.py

@@ -62,6 +62,7 @@ def get_file_text(file_name):
         "httpx": ["httpx>=0.16.0"],
         "starlette": ["starlette>=0.19.1"],
         "fastapi": ["fastapi>=0.79.0"],
+        "pymongo": ["pymongo>=3.1"],
     },
     classifiers=[
         "Development Status :: 5 - Production/Stable",

tests/integrations/pymongo/__init__.py

@@ -0,0 +1,3 @@
+import pytest
+
+pytest.importorskip("pymongo")