Optional Anonymous Telemetry (#727)

backend/danswer/auth/users.py

@@ -48,6 +48,8 @@
 from danswer.db.models import AccessToken
 from danswer.db.models import User
 from danswer.utils.logger import setup_logger
+from danswer.utils.telemetry import optional_telemetry
+from danswer.utils.telemetry import RecordType
 from danswer.utils.variable_functionality import fetch_versioned_implementation
 
 
@@ -170,6 +172,7 @@ async def on_after_register(
         self, user: User, request: Optional[Request] = None
     ) -> None:
         logger.info(f"User {user.id} has registered.")
+        optional_telemetry(record_type=RecordType.SIGN_UP, data={"user": "create"})
 
     async def on_after_forgot_password(
         self, user: User, token: str, request: Optional[Request] = None

backend/danswer/configs/app_configs.py

@@ -238,5 +238,7 @@
 LOG_ALL_MODEL_INTERACTIONS = (
     os.environ.get("LOG_ALL_MODEL_INTERACTIONS", "").lower() == "true"
 )
+# Anonymous usage telemetry
+DISABLE_TELEMETRY = os.environ.get("DISABLE_TELEMETRY", "").lower() == "true"
 # notset, debug, info, warning, error, or critical
 LOG_LEVEL = os.environ.get("LOG_LEVEL", "info")

backend/danswer/db/index_attempt.py

@@ -13,7 +13,8 @@
 from danswer.db.models import IndexingStatus
 from danswer.server.models import ConnectorCredentialPairIdentifier
 from danswer.utils.logger import setup_logger
-
+from danswer.utils.telemetry import optional_telemetry
+from danswer.utils.telemetry import RecordType
 
 logger = setup_logger()
 
@@ -88,6 +89,9 @@ def mark_attempt_failed(
     db_session.add(index_attempt)
     db_session.commit()
 
+    source = index_attempt.connector.source
+    optional_telemetry(record_type=RecordType.FAILURE, data={"connector": source})
+
 
 def update_docs_indexed(
     db_session: Session,

backend/danswer/main.py

@@ -48,6 +48,8 @@
 from danswer.server.state import router as state_router
 from danswer.server.users import router as user_router
 from danswer.utils.logger import setup_logger
+from danswer.utils.telemetry import optional_telemetry
+from danswer.utils.telemetry import RecordType
 from danswer.utils.variable_functionality import fetch_versioned_implementation
 
 
@@ -212,6 +214,10 @@ def startup_event() -> None:
         logger.info("Verifying Document Index(s) is/are available.")
         get_default_document_index().ensure_indices_exist()
 
+        optional_telemetry(
+            record_type=RecordType.VERSION, data={"version": __version__}
+        )
+
     application.add_middleware(
         CORSMiddleware,
         allow_origins=["*"],  # Change this to the list of allowed origins if needed

backend/danswer/utils/telemetry.py

@@ -0,0 +1,56 @@
+import threading
+import uuid
+from enum import Enum
+from typing import cast
+
+import requests
+
+from danswer.configs.app_configs import DISABLE_TELEMETRY
+from danswer.dynamic_configs import get_dynamic_config_store
+from danswer.dynamic_configs.interface import ConfigNotFoundError
+
+CUSTOMER_UUID_KEY = "customer_uuid"
+DANSWER_TELEMETRY_ENDPOINT = "https://telemetry.danswer.ai/anonymous_telemetry"
+
+
+class RecordType(str, Enum):
+    VERSION = "version"
+    SIGN_UP = "sign_up"
+    LATENCY = "latency"
+    FAILURE = "failure"
+
+
+def get_or_generate_uuid() -> str:
+    kv_store = get_dynamic_config_store()
+    try:
+        return cast(str, kv_store.load(CUSTOMER_UUID_KEY))
+    except ConfigNotFoundError:
+        customer_id = str(uuid.uuid4())
+        kv_store.store(CUSTOMER_UUID_KEY, customer_id)
+        return customer_id
+
+
+def optional_telemetry(record_type: RecordType, data: dict) -> None:
+    if DISABLE_TELEMETRY:
+        return
+
+    try:
+
+        def telemetry_logic() -> None:
+            payload = {
+                "data": data,
+                "record": record_type,
+                "customer_uuid": get_or_generate_uuid(),
+            }
+            requests.post(
+                DANSWER_TELEMETRY_ENDPOINT,
+                headers={"Content-Type": "application/json"},
+                json=payload,
+            )
+
+        # Run in separate thread to have minimal overhead in main flows
+        thread = threading.Thread(target=telemetry_logic, daemon=True)
+        thread.start()
+    except Exception:
+        # Should never interfere with normal functions of Danswer
+        pass

backend/danswer/utils/timing.py

@@ -8,6 +8,8 @@
 from typing import TypeVar
 
 from danswer.utils.logger import setup_logger
+from danswer.utils.telemetry import optional_telemetry
+from danswer.utils.telemetry import RecordType
 
 logger = setup_logger()
 
@@ -21,8 +23,13 @@ def decorator(func: F) -> F:
         def wrapped_func(*args: Any, **kwargs: Any) -> Any:
             start_time = time.time()
             result = func(*args, **kwargs)
-            elapsed_time = time.time() - start_time
-            logger.info(f"{func_name or func.__name__} took {elapsed_time} seconds")
+            elapsed_time_str = str(time.time() - start_time)
+            log_name = func_name or func.__name__
+            logger.info(f"{log_name} took {elapsed_time_str} seconds")
+            optional_telemetry(
+                record_type=RecordType.LATENCY,
+                data={"function": log_name, "latency": str(elapsed_time_str)},
+            )
             return result
 
         return cast(F, wrapped_func)
@@ -44,8 +51,13 @@ def wrapped_func(*args: Any, **kwargs: Any) -> Any:
             except StopIteration:
                 pass
             finally:
-                elapsed_time = time.time() - start_time
-                logger.info(f"{func_name or func.__name__} took {elapsed_time} seconds")
+                elapsed_time_str = str(time.time() - start_time)
+                log_name = func_name or func.__name__
+                logger.info(f"{log_name} took {elapsed_time_str} seconds")
+                optional_telemetry(
+                    record_type=RecordType.LATENCY,
+                    data={"function": log_name, "latency": str(elapsed_time_str)},
+                )
 
         return cast(FG, wrapped_func)
 

deployment/docker_compose/docker-compose.dev.yml

@@ -48,6 +48,9 @@ services:
       - EDIT_KEYWORD_QUERY=${EDIT_KEYWORD_QUERY:-}
       - MODEL_SERVER_HOST=${MODEL_SERVER_HOST:-}
       - MODEL_SERVER_PORT=${MODEL_SERVER_PORT:-}
+      # Leave this on pretty please? Nothing sensitive is collected!
+      # https://docs.danswer.dev/more/telemetry
+      - DISABLE_TELEMETRY=${DISABLE_TELEMETRY:-}
       # Set to debug to get more fine-grained logs
       - LOG_LEVEL=${LOG_LEVEL:-info}
     volumes: