elasticsearch: don't set body as db statement for bulk requests (#2355)

* elasticsearch: don't set body as db statement for bulk requests

bulk requests can be too big and diverse to make sense as db statement.
Other than that the sanitizer currently only handles dicts so it's
crashing.

Closes #2150

Co-authored-by: Jason Mobarak <[email protected]>
Co-authored-by: Quentin Pradet <[email protected]>

* Update CHANGELOG

* Please the linter

---------

Co-authored-by: Jason Mobarak <[email protected]>
Co-authored-by: Quentin Pradet <[email protected]>

Author: 3 people

CHANGELOG.md

@@ -21,6 +21,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
   ([#2151](https://github.com/open-telemetry/opentelemetry-python-contrib/issues/2298))
 - Avoid losing repeated HTTP headers
   ([#2266](https://github.com/open-telemetry/opentelemetry-python-contrib/pull/2266))
+- `opentelemetry-instrumentation-elasticsearch` Don't send bulk request body as db statement
+  ([#2355](https://github.com/open-telemetry/opentelemetry-python-contrib/pull/2355))
 
 ## Version 1.23.0/0.44b0 (2024-02-23)
 

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

@@ -245,9 +245,11 @@ def wrapper(wrapped, _, args, kwargs):
                 if method:
                     attributes["elasticsearch.method"] = method
                 if body:
-                    attributes[SpanAttributes.DB_STATEMENT] = sanitize_body(
-                        body
-                    )
+                    # Don't set db.statement for bulk requests, as it can be very large
+                    if isinstance(body, dict):
+                        attributes[
+                            SpanAttributes.DB_STATEMENT
+                        ] = sanitize_body(body)
                 if params:
                     attributes["elasticsearch.params"] = str(params)
                 if doc_id:

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

@@ -51,6 +51,8 @@
 
 Article = helpers.Article
 
+# pylint: disable=too-many-public-methods
+
 
 @mock.patch(
     "elasticsearch.connection.http_urllib3.Urllib3HttpConnection.perform_request"
@@ -486,3 +488,35 @@ def test_body_sanitization(self, _):
             sanitize_body(json.dumps(sanitization_queries.interval_query)),
             str(sanitization_queries.interval_query_sanitized),
         )
+
+    def test_bulk(self, request_mock):
+        request_mock.return_value = (1, {}, "")
+
+        es = Elasticsearch()
+        es.bulk(
+            [
+                {
+                    "_op_type": "index",
+                    "_index": "sw",
+                    "_doc_type": "_doc",
+                    "_id": 1,
+                    "doc": {"name": "adam"},
+                },
+                {
+                    "_op_type": "index",
+                    "_index": "sw",
+                    "_doc_type": "_doc",
+                    "_id": 1,
+                    "doc": {"name": "adam"},
+                },
+            ]
+        )
+
+        spans_list = self.get_finished_spans()
+        self.assertEqual(len(spans_list), 1)
+        span = spans_list[0]
+
+        # Check version and name in span's instrumentation info
+        self.assertEqualSpanInstrumentationInfo(
+            span, opentelemetry.instrumentation.elasticsearch
+        )