add elasticsearch db.statement sanitization

CHANGELOG.md

@@ -9,6 +9,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### Added
 
+- Add default query sanitization for elasticsearch db.statement attribute
+  ([#1545](https://github.com/open-telemetry/opentelemetry-python-contrib/pull/1545))
 - Add metric instrumentation for urllib
   ([#1553](https://github.com/open-telemetry/opentelemetry-python-contrib/pull/1553))
 - `opentelemetry/sdk/extension/aws` Implement [`aws.ecs.*`](https://github.com/open-telemetry/opentelemetry-specification/blob/main/specification/resource/semantic_conventions/cloud_provider/aws/ecs.md) and [`aws.logs.*`](https://opentelemetry.io/docs/reference/specification/resource/semantic_conventions/cloud_provider/aws/logs/) resource attributes in the `AwsEcsResourceDetector` detector when the ECS Metadata v4 is available

instrumentation/opentelemetry-instrumentation-elasticsearch/src/opentelemetry/instrumentation/elasticsearch/__init__.py

@@ -135,11 +135,16 @@ def _instrument(self, **kwargs):
         tracer = get_tracer(__name__, __version__, tracer_provider)
         request_hook = kwargs.get("request_hook")
         response_hook = kwargs.get("response_hook")
+        sanitize_query = kwargs.get("sanitize_query", True)
         _wrap(
             elasticsearch,
             "Transport.perform_request",
             _wrap_perform_request(
-                tracer, self._span_name_prefix, request_hook, response_hook
+                tracer,
+                sanitize_query,
+                self._span_name_prefix,
+                request_hook,
+                response_hook,
             ),
         )
 
@@ -154,7 +159,11 @@ def _uninstrument(self, **kwargs):
 
 
 def _wrap_perform_request(
-    tracer, span_name_prefix, request_hook=None, response_hook=None
+    tracer,
+    sanitize_query,
+    span_name_prefix,
+    request_hook=None,
+    response_hook=None,
 ):
     # pylint: disable=R0912,R0914
     def wrapper(wrapped, _, args, kwargs):
@@ -214,7 +223,10 @@ def wrapper(wrapped, _, args, kwargs):
                 if method:
                     attributes["elasticsearch.method"] = method
                 if body:
-                    attributes[SpanAttributes.DB_STATEMENT] = str(body)
+                    statement = str(body)
+                    if sanitize_query:
+                        statement = "sanitized"
+                    attributes[SpanAttributes.DB_STATEMENT] = statement
                 if params:
                     attributes["elasticsearch.params"] = str(params)
                 if doc_id:

instrumentation/opentelemetry-instrumentation-elasticsearch/tests/test_elasticsearch.py

@@ -42,18 +42,33 @@
 else:
     from . import helpers_es2 as helpers  # pylint: disable=no-name-in-module
 
-
 Article = helpers.Article
 
 
 @mock.patch(
     "elasticsearch.connection.http_urllib3.Urllib3HttpConnection.perform_request"
 )
 class TestElasticsearchIntegration(TestBase):
+    search_attributes = {
+        SpanAttributes.DB_SYSTEM: "elasticsearch",
+        "elasticsearch.url": "/test-index/_search",
+        "elasticsearch.method": helpers.dsl_search_method,
+        "elasticsearch.target": "test-index",
+        SpanAttributes.DB_STATEMENT: str(
+            {"query": {"bool": {"filter": [{"term": {"author": "testing"}}]}}}
+        ),
+    }
+
+    create_attributes = {
+        SpanAttributes.DB_SYSTEM: "elasticsearch",
+        "elasticsearch.url": "/test-index",
+        "elasticsearch.method": "HEAD",
+    }
+
     def setUp(self):
         super().setUp()
         self.tracer = self.tracer_provider.get_tracer(__name__)
-        ElasticsearchInstrumentor().instrument()
+        ElasticsearchInstrumentor().instrument(sanitize_query=False)
 
     def tearDown(self):
         super().tearDown()
@@ -241,21 +256,34 @@ def test_dsl_search(self, request_mock):
         self.assertIsNotNone(span.end_time)
         self.assertEqual(
             span.attributes,
-            {
-                SpanAttributes.DB_SYSTEM: "elasticsearch",
-                "elasticsearch.url": "/test-index/_search",
-                "elasticsearch.method": helpers.dsl_search_method,
-                "elasticsearch.target": "test-index",
-                SpanAttributes.DB_STATEMENT: str(
-                    {
-                        "query": {
-                            "bool": {
-                                "filter": [{"term": {"author": "testing"}}]
-                            }
-                        }
-                    }
-                ),
-            },
+            self.search_attributes,
+        )
+
+    def test_dsl_search_sanitized(self, request_mock):
+        # Reset instrumentation to use sanitized query (default)
+        ElasticsearchInstrumentor().uninstrument()
+        ElasticsearchInstrumentor().instrument()
+
+        # update expected attributes to match sanitized query
+        sanitized_search_attributes = self.search_attributes.copy()
+        sanitized_search_attributes.update(
+            {SpanAttributes.DB_STATEMENT: "sanitized"}
+        )
+
+        request_mock.return_value = (1, {}, '{"hits": {"hits": []}}')
+        client = Elasticsearch()
+        search = Search(using=client, index="test-index").filter(
+            "term", author="testing"
+        )
+        search.execute()
+        spans = self.get_finished_spans()
+        span = spans[0]
+        self.assertEqual(1, len(spans))
+        self.assertEqual(span.name, "Elasticsearch/<target>/_search")
+        self.assertIsNotNone(span.end_time)
+        self.assertEqual(
+            span.attributes,
+            sanitized_search_attributes,
         )
 
     def test_dsl_create(self, request_mock):
@@ -270,11 +298,7 @@ def test_dsl_create(self, request_mock):
 
         self.assertEqual(
             span1.attributes,
-            {
-                SpanAttributes.DB_SYSTEM: "elasticsearch",
-                "elasticsearch.url": "/test-index",
-                "elasticsearch.method": "HEAD",
-            },
+            self.create_attributes,
         )
 
         attributes = {
@@ -288,6 +312,24 @@ def test_dsl_create(self, request_mock):
             helpers.dsl_create_statement,
         )
 
+    def test_dsl_create_sanitized(self, request_mock):
+        # Reset instrumentation to explicitly use sanitized query
+        ElasticsearchInstrumentor().uninstrument()
+        ElasticsearchInstrumentor().instrument(sanitize_query=True)
+
+        request_mock.return_value = (1, {}, {})
+        client = Elasticsearch()
+        Article.init(using=client)
+
+        spans = self.get_finished_spans()
+        self.assertEqual(2, len(spans))
+        span = spans.by_attr(key="elasticsearch.method", value="HEAD")
+
+        self.assertEqual(
+            span.attributes,
+            self.create_attributes,
+        )
+
     def test_dsl_index(self, request_mock):
         request_mock.return_value = helpers.dsl_index_result
 
@@ -323,7 +365,6 @@ def test_request_hook(self, request_mock):
         request_hook_kwargs_attribute = "request_hook.kwargs"
 
         def request_hook(span, method, url, kwargs):
-
             attributes = {
                 request_hook_method_attribute: method,
                 request_hook_url_attribute: url,