parse apigw data as chained trigger (#270)

* parse apigw data as chained trigger
* use the event_id as session hash

Author: saartochner-lumigo

.secrets.baseline

@@ -168,14 +168,14 @@
         "filename": "src/test/unit/event/test_event_trigger.py",
         "hashed_secret": "885bb9903f72e004ff2974807b70e7c970d3e6d5",
         "is_verified": false,
-        "line_number": 612
+        "line_number": 613
       },
       {
         "type": "Hex High Entropy String",
         "filename": "src/test/unit/event/test_event_trigger.py",
         "hashed_secret": "3fae06dc55a618caed1d794dcd512bfe7e76c9f1",
         "is_verified": false,
-        "line_number": 657
+        "line_number": 658
       }
     ],
     "src/test/unit/test_lumigo_utils.py": [
@@ -220,5 +220,5 @@
       }
     ]
   },
-  "generated_at": "2023-01-22T10:15:06Z"
+  "generated_at": "2023-01-29T12:29:33Z"
 }

src/lumigo_tracer/event/event_trigger.py

@@ -19,7 +19,7 @@ def recursive_parse_trigger(
             triggers.append(new_trigger)
             current_trigger_id: str = new_trigger["id"]  # type: ignore
 
-            inner_messages = parser.extract_inner(event=message)
+            inner_messages = parser.extract_inner_message(event=message)
             if len(inner_messages) >= Configuration.chained_services_max_width:
                 get_logger().info("Chained services parsing has stopped due to width")
                 inner_messages = inner_messages[: Configuration.chained_services_max_width]
@@ -31,6 +31,9 @@ def recursive_parse_trigger(
                             json.loads(sub_message), parent_id=current_trigger_id, level=level + 1
                         )
                     )
+            inner_triggers = parser.extract_inner_triggers(message, current_trigger_id)
+            for inner_trigger in inner_triggers:
+                triggers.append(inner_trigger)
             break
     return triggers
 

src/lumigo_tracer/event/trigger_parsing/apigw_parser.py

@@ -1,10 +1,11 @@
-from typing import Optional, Dict, Any
+from typing import Optional, Dict, Any, List
 
 from lumigo_tracer.event.trigger_parsing.event_trigger_base import (
     EventTriggerParser,
     ExtraKeys,
     TriggerType,
 )
+from lumigo_tracer.lumigo_utils import get_current_ms_time
 
 
 class ApiGatewayEventTriggerParser(EventTriggerParser):
@@ -56,3 +57,26 @@ def _parse_http_method_v2(event: Dict[Any, Any], target_id: Optional[str]) -> Tr
             from_message_ids=[event.get("requestContext", {}).get("requestId", "")],
             extra=extra,
         )
+
+    @staticmethod
+    def extract_inner_triggers(event: Dict[Any, Any], target_id: str) -> List[TriggerType]:
+        """
+        In this function we extract the browser's session using the authorization ID:
+        https://openid.net/specs/openid-connect-core-1_0.html#CodeIDToken
+        """
+        user_agent = event.get("headers", {}).get("User-Agent")
+        claims = event.get("requestContext", {}).get("authorizer", {}).get("claims", {})
+        auth_hash = claims.get("at_hash") or claims.get("event_id")
+        if user_agent and auth_hash:
+            return [
+                EventTriggerParser.build_trigger(
+                    target_id=target_id,
+                    resource_type="browser",
+                    from_message_ids=[auth_hash],
+                    extra={
+                        ExtraKeys.USER_AGENT: user_agent,
+                        ExtraKeys.TRIGGER_CREATION_TIME: get_current_ms_time(),
+                    },
+                )
+            ]
+        return []

src/lumigo_tracer/event/trigger_parsing/event_trigger_base.py

@@ -16,6 +16,7 @@ class ExtraKeys(str, Enum):
     TRIGGER_CREATION_TIME = "approxEventCreationTime"
     TOTAL_SIZE = "totalSizeBytes"
     SHARD_ID = "shardId"
+    USER_AGENT = "userAgent"
 
 
 ExtraType = Dict[ExtraKeys, Union[str, int, None]]
@@ -63,5 +64,9 @@ def handle(event: Dict[Any, Any], target_id: Optional[str]) -> TriggerType:
         raise NotImplementedError()
 
     @staticmethod
-    def extract_inner(event: Dict[Any, Any]) -> List[str]:
+    def extract_inner_triggers(event: Dict[Any, Any], target_id: str) -> List[TriggerType]:
+        return []
+
+    @staticmethod
+    def extract_inner_message(event: Dict[Any, Any]) -> List[str]:
         return []

src/lumigo_tracer/event/trigger_parsing/sqs_parser.py

@@ -40,7 +40,7 @@ def handle(event: Dict[Any, Any], target_id: Optional[str]) -> TriggerType:
         )
 
     @staticmethod
-    def extract_inner(event: Dict[Any, Any]) -> List[str]:
+    def extract_inner_message(event: Dict[Any, Any]) -> List[str]:
         inner_messages = []
         for record in SqsEventTriggerParser._get_messages(event):
             body = record.get("body") or record.get("Body")

src/test/unit/event/test_event_trigger.py

@@ -4,6 +4,7 @@
 
 from lumigo_tracer.event.event_trigger import parse_triggers
 from lumigo_tracer.event.trigger_parsing import INNER_MESSAGES_MAGIC_PATTERN
+from lumigo_tracer.event.trigger_parsing.event_trigger_base import ExtraKeys
 from lumigo_tracer.lumigo_utils import Configuration
 
 
@@ -884,3 +885,33 @@ def test_recursive_triggers_too_wide(caplog):
     assert any(
         "Chained services parsing has stopped due to width" in log.message for log in caplog.records
     )
+
+
+def test_apigw_session():
+    user_agent = "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/109.0.0.0 Safari/537.36"
+    apigw_event = {
+        "resource": "/projects/resource",
+        "path": "/projects/path",
+        "httpMethod": "POST",
+        "pathParameters": {},
+        "body": "",
+        "requestContext": {
+            "stage": "dev",
+            "authorizer": {
+                "claims": {
+                    "at_hash": "at_hash",
+                }
+            },
+            "requestId": "aaaa-bbbb-cccc-ddddd",
+        },
+        "headers": {"User-Agent": user_agent},
+        "version": "2.0",
+    }
+    triggers = parse_triggers(apigw_event)
+    assert len(triggers) == 2
+    apigw = [t for t in triggers if t["triggeredBy"] == "apigw"][0]
+    browser = [t for t in triggers if t["triggeredBy"] == "browser"][0]
+    assert apigw["id"] == browser["targetId"]
+    assert browser["fromMessageIds"] == ["at_hash"]
+    assert browser["extra"][ExtraKeys.USER_AGENT] == user_agent
+    assert browser["extra"][ExtraKeys.TRIGGER_CREATION_TIME]