feat(sqlalchemy): Support span streaming

sentry_sdk/integrations/asyncpg.py

@@ -1,7 +1,7 @@
 from __future__ import annotations
 import contextlib
 import re
-from typing import Any, TypeVar, Callable, Awaitable, Iterator
+from typing import Any, TypeVar, Callable, Awaitable, Iterator, Union
 
 import sentry_sdk
 from sentry_sdk.consts import OP, SPANDATA
@@ -13,6 +13,7 @@
     parse_version,
     capture_internal_exceptions,
 )
+from sentry_sdk.traces import StreamedSpan
 
 try:
     import asyncpg  # type: ignore[import-not-found]
@@ -101,7 +102,7 @@
     params_list: "tuple[Any, ...] | None",
     *,
     executemany: bool = False,
-) -> "Iterator[Span]":
+) -> "Iterator[Union[Span, StreamedSpan]]":
     integration = sentry_sdk.get_client().get_integration(AsyncPGIntegration)
     if integration is not None and not integration._record_params:
         params_list = None
@@ -197,22 +198,27 @@
     return _inner
 
 
-def _set_db_data(span: "Span", conn: "Any") -> None:
-    span.set_data(SPANDATA.DB_SYSTEM, "postgresql")
-    span.set_data(SPANDATA.DB_DRIVER_NAME, "asyncpg")
+def _set_db_data(span: "Union[Span, StreamedSpan]", conn: "Any") -> None:
+    if isinstance(span, StreamedSpan):
+        set_on_span = span.set_attribute
+    else:
+        set_on_span = span.set_data
+
+    set_on_span(SPANDATA.DB_SYSTEM, "postgresql")
+    set_on_span(SPANDATA.DB_DRIVER_NAME, "asyncpg")
 
     addr = conn._addr
     if addr:
         try:
-            span.set_data(SPANDATA.SERVER_ADDRESS, addr[0])
-            span.set_data(SPANDATA.SERVER_PORT, addr[1])
+            set_on_span(SPANDATA.SERVER_ADDRESS, addr[0])
+            set_on_span(SPANDATA.SERVER_PORT, addr[1])
         except IndexError:
             pass
 
     database = conn._params.database
     if database:
-        span.set_data(SPANDATA.DB_NAME, database)
+        set_on_span(SPANDATA.DB_NAME, database)
 
     user = conn._params.user
     if user:
-        span.set_data(SPANDATA.DB_USER, user)
+        set_on_span(SPANDATA.DB_USER, user)

sentry_sdk/integrations/django/__init__.py

@@ -9,6 +9,7 @@
 from sentry_sdk.scope import add_global_event_processor, should_send_default_pii
 from sentry_sdk.serializer import add_global_repr_processor, add_repr_sequence_type
 from sentry_sdk.tracing import SOURCE_FOR_STYLE, TransactionSource
+from sentry_sdk.traces import StreamedSpan
 from sentry_sdk.tracing_utils import add_query_source, record_sql_queries
 from sentry_sdk.utils import (
     AnnotatedValue,
@@ -720,14 +721,21 @@ def _rollback(self: "BaseDatabaseWrapper") -> None:
 
 
 def _set_db_data(
-    span: "Span", cursor_or_db: "Any", db_operation: "Optional[str]" = None
+    span: "Union[Span, StreamedSpan]",
+    cursor_or_db: "Any",
+    db_operation: "Optional[str]" = None,
 ) -> None:
+    if isinstance(span, StreamedSpan):
+        set_on_span = span.set_attribute
+    else:
+        set_on_span = span.set_data
+
     db = cursor_or_db.db if hasattr(cursor_or_db, "db") else cursor_or_db
     vendor = db.vendor
-    span.set_data(SPANDATA.DB_SYSTEM, vendor)
+    set_on_span(SPANDATA.DB_SYSTEM, vendor)
 
     if db_operation is not None:
-        span.set_data(SPANDATA.DB_OPERATION, db_operation)
+        set_on_span(SPANDATA.DB_OPERATION, db_operation)
 
     # Some custom backends override `__getattr__`, making it look like `cursor_or_db`
     # actually has a `connection` and the `connection` has a `get_dsn_parameters`
@@ -760,19 +768,19 @@ def _set_db_data(
 
     db_name = connection_params.get("dbname") or connection_params.get("database")
     if db_name is not None:
-        span.set_data(SPANDATA.DB_NAME, db_name)
+        set_on_span(SPANDATA.DB_NAME, db_name)
 
     server_address = connection_params.get("host")
     if server_address is not None:
-        span.set_data(SPANDATA.SERVER_ADDRESS, server_address)
+        set_on_span(SPANDATA.SERVER_ADDRESS, server_address)
 
     server_port = connection_params.get("port")
     if server_port is not None:
-        span.set_data(SPANDATA.SERVER_PORT, str(server_port))
+        set_on_span(SPANDATA.SERVER_PORT, str(server_port))
 
     server_socket_address = connection_params.get("unix_socket")
     if server_socket_address is not None:
-        span.set_data(SPANDATA.SERVER_SOCKET_ADDRESS, server_socket_address)
+        set_on_span(SPANDATA.SERVER_SOCKET_ADDRESS, server_socket_address)
 
 
 def add_template_context_repr_sequence() -> None:

sentry_sdk/integrations/sqlalchemy.py

@@ -6,6 +6,7 @@
     ensure_integration_enabled,
     parse_version,
 )
+from sentry_sdk.traces import StreamedSpan, SpanStatus
 
 try:
     from sqlalchemy.engine import Engine  # type: ignore
@@ -20,6 +21,7 @@
     from typing import Any
     from typing import ContextManager
     from typing import Optional
+    from typing import Union
 
     from sentry_sdk.tracing import Span
 
@@ -96,7 +98,10 @@ def _handle_error(context: "Any", *args: "Any") -> None:
     span: "Optional[Span]" = getattr(execution_context, "_sentry_sql_span", None)
 
     if span is not None:
-        span.set_status(SPANSTATUS.INTERNAL_ERROR)
+        if isinstance(span, StreamedSpan):
+            span.status = SpanStatus.ERROR
+        else:
+            span.set_status(SPANSTATUS.INTERNAL_ERROR)
 
     # _after_cursor_execute does not get called for crashing SQL stmts. Judging
     # from SQLAlchemy codebase it does seem like any error coming into this
@@ -132,15 +137,20 @@ def _get_db_system(name: str) -> "Optional[str]":
     return None
 
 
-def _set_db_data(span: "Span", conn: "Any") -> None:
+def _set_db_data(span: "Union[Span, StreamedSpan]", conn: "Any") -> None:
+    if isinstance(span, StreamedSpan):
+        set_on_span = span.set_attribute
+    else:
+        set_on_span = span.set_data
+
     db_system = _get_db_system(conn.engine.name)
     if db_system is not None:
-        span.set_data(SPANDATA.DB_SYSTEM, db_system)
+        set_on_span(SPANDATA.DB_SYSTEM, db_system)
 
     try:
         driver = conn.dialect.driver
         if driver:
-            span.set_data(SPANDATA.DB_DRIVER_NAME, driver)
+            set_on_span(SPANDATA.DB_DRIVER_NAME, driver)
     except Exception:
         pass
 
@@ -149,12 +159,12 @@ def _set_db_data(span: "Span", conn: "Any") -> None:
 
     db_name = conn.engine.url.database
     if db_name is not None:
-        span.set_data(SPANDATA.DB_NAME, db_name)
+        set_on_span(SPANDATA.DB_NAME, db_name)
 
     server_address = conn.engine.url.host
     if server_address is not None:
-        span.set_data(SPANDATA.SERVER_ADDRESS, server_address)
+        set_on_span(SPANDATA.SERVER_ADDRESS, server_address)
 
     server_port = conn.engine.url.port
     if server_port is not None:
-        span.set_data(SPANDATA.SERVER_PORT, server_port)
+        set_on_span(SPANDATA.SERVER_PORT, server_port)

sentry_sdk/tracing_utils.py

@@ -133,9 +133,10 @@ def record_sql_queries(
     executemany: bool,
     record_cursor_repr: bool = False,
     span_origin: str = "manual",
-) -> "Generator[sentry_sdk.tracing.Span, None, None]":
+) -> "Generator[Union[sentry_sdk.tracing.Span, sentry_sdk.traces.StreamedSpan], None, None]":
     # TODO: Bring back capturing of params by default
-    if sentry_sdk.get_client().options["_experiments"].get("record_sql_params", False):
+    client = sentry_sdk.get_client()
+    if client.options["_experiments"].get("record_sql_params", False):
         if not params_list or params_list == [None]:
             params_list = None
 
@@ -160,14 +161,26 @@ def record_sql_queries(
     with capture_internal_exceptions():
         sentry_sdk.add_breadcrumb(message=query, category="query", data=data)
 
-    with sentry_sdk.start_span(
-        op=OP.DB,
-        name=query,
-        origin=span_origin,
-    ) as span:
-        for k, v in data.items():
-            span.set_data(k, v)
-        yield span
+    if has_span_streaming_enabled(client.options):
+        with sentry_sdk.traces.start_span(
+            name="<unknown SQL query>" if query is None else query,
+            attributes={
+                "sentry.origin": span_origin,
+                "sentry.op": OP.DB,
+            },
+        ) as span:
+            for k, v in data.items():
+                span.set_attribute(k, v)
+            yield span
+    else:
+        with sentry_sdk.start_span(
+            op=OP.DB,
+            name=query,
+            origin=span_origin,
+        ) as span:
+            for k, v in data.items():
+                span.set_data(k, v)
+            yield span
 
 
 def maybe_create_breadcrumbs_from_span(
@@ -313,22 +326,36 @@ def add_source(
                 span.set_attribute("code.function.name", frame.f_code.co_name)
 
 
-def add_query_source(span: "sentry_sdk.tracing.Span") -> None:
+def add_query_source(
+    span: "Union[sentry_sdk.tracing.Span, sentry_sdk.traces.StreamedSpan]",
+) -> None:
     """
     Adds OTel compatible source code information to a database query span
     """
     client = sentry_sdk.get_client()
-    if not client.is_active():
-        return
 
-    if span.timestamp is None or span.start_timestamp is None:
+    if isinstance(span, LegacySpan):
+        if not client.is_active():
+            return
+
+        # In the StreamedSpan case, we need to add the extra span information before
+        # the span finishes, so it's expected that this will be None. In the LegacySpan case,
+        # it should already be finished.
+        if span.timestamp is None:
+            return
+
+    if span.start_timestamp is None:
         return
 
     should_add_query_source = client.options.get("enable_db_query_source", True)
     if not should_add_query_source:
         return
 
-    duration = span.timestamp - span.start_timestamp
+    end_timestamp = (
+        datetime.now(timezone.utc) if span.timestamp is None else span.timestamp
+    )
+
+    duration = end_timestamp - span.start_timestamp
     threshold = client.options.get("db_query_source_threshold_ms", 0)
     slow_query = duration / timedelta(milliseconds=1) > threshold
 

tests/conftest.py

@@ -476,24 +476,36 @@
 
 @pytest.fixture
 def render_span_tree():
-    def inner(event):
-        assert event["type"] == "transaction"
+    def inner(spans, root_span=None):
+        streamed_spans = False
+        if root_span is None:
+            streamed_spans = True
 
         by_parent = {}
-        for span in event["spans"]:
+        for span in spans:
+            print(span)
+            if "parent_span_id" not in span:
+                root_span = span
+                continue
+
             by_parent.setdefault(span["parent_span_id"], []).append(span)
 
         def render_span(span):
-            yield "- op={}: description={}".format(
-                json.dumps(span.get("op")), json.dumps(span.get("description"))
-            )
+            if streamed_spans:
+                yield "- sentry.op={}: name={}".format(
+                    json.dumps(span["attributes"].get("sentry.op")),
+                    json.dumps(span["name"]),
+                )
+            else:
+                yield "- op={}: description={}".format(
+                    json.dumps(span.get("op")), json.dumps(span.get("description"))
+                )
+
             for subspan in by_parent.get(span["span_id"]) or ():
                 for line in render_span(subspan):
                     yield "  {}".format(line)
 
-        root_span = event["contexts"]["trace"]
-
         return "\n".join(render_span(root_span))
 
     return inner
 

tests/integrations/django/asgi/test_asgi.py

@@ -292,8 +292,9 @@ async def test_async_middleware_spans(
 
     (transaction,) = events
 
+    assert transaction["type"] == "transaction"
     assert (
-        render_span_tree(transaction)
+        render_span_tree(transaction["spans"], transaction["contexts"]["trace"])
         == """\
 - op="http.server": description=null
   - op="event.django": description="django.db.reset_queries"