Setting HTTP timeout to just 5 seconds, so we can fail fast if there'…

…s a networking error.

src/groundlight/client.py

@@ -28,6 +28,11 @@
 
 logger = logging.getLogger("groundlight.sdk")
 
+# Set urllib3 request timeout to something modern and fast.
+# The system defaults can be stupidly long
+# It used to take >8 min to timeout to a bad IP address
+DEFAULT_REQUEST_TIMEOUT = 5
+
 
 class ApiTokenError(Exception):
     pass
@@ -154,7 +159,7 @@ def get_detector(self, id: Union[str, Detector]) -> Detector:  # pylint: disable
         if isinstance(id, Detector):
             # Short-circuit
             return id
-        obj = self.detectors_api.get_detector(id=id)
+        obj = self.detectors_api.get_detector(id=id, _request_timeout=DEFAULT_REQUEST_TIMEOUT)
         return Detector.parse_obj(obj.to_dict())
 
     def get_detector_by_name(self, name: str) -> Detector:
@@ -177,7 +182,9 @@ def list_detectors(self, page: int = 1, page_size: int = 10) -> PaginatedDetecto
 
         :return: PaginatedDetectorList
         """
-        obj = self.detectors_api.list_detectors(page=page, page_size=page_size)
+        obj = self.detectors_api.list_detectors(
+            page=page, page_size=page_size, _request_timeout=DEFAULT_REQUEST_TIMEOUT
+        )
         return PaginatedDetectorList.parse_obj(obj.to_dict())
 
     def create_detector(
@@ -213,7 +220,7 @@ def create_detector(
             detector_creation_input.pipeline_config = pipeline_config
         if metadata is not None:
             detector_creation_input.metadata = str(url_encode_dict(metadata, name="metadata", size_limit_bytes=1024))
-        obj = self.detectors_api.create_detector(detector_creation_input)
+        obj = self.detectors_api.create_detector(detector_creation_input, _request_timeout=DEFAULT_REQUEST_TIMEOUT)
         return Detector.parse_obj(obj.to_dict())
 
     def get_or_create_detector(
@@ -282,7 +289,7 @@ def get_image_query(self, id: str) -> ImageQuery:  # pylint: disable=redefined-b
 
         :return: ImageQuery
         """
-        obj = self.image_queries_api.get_image_query(id=id)
+        obj = self.image_queries_api.get_image_query(id=id, _request_timeout=DEFAULT_REQUEST_TIMEOUT)
         iq = ImageQuery.parse_obj(obj.to_dict())
         return self._fixup_image_query(iq)
 
@@ -296,7 +303,9 @@ def list_image_queries(self, page: int = 1, page_size: int = 10) -> PaginatedIma
 
         :return: PaginatedImageQueryList
         """
-        obj = self.image_queries_api.list_image_queries(page=page, page_size=page_size)
+        obj = self.image_queries_api.list_image_queries(
+            page=page, page_size=page_size, _request_timeout=DEFAULT_REQUEST_TIMEOUT
+        )
         image_queries = PaginatedImageQueryList.parse_obj(obj.to_dict())
         if image_queries.results is not None:
             image_queries.results = [self._fixup_image_query(iq) for iq in image_queries.results]
@@ -385,6 +394,8 @@ def submit_image_query(  # noqa: PLR0913 # pylint: disable=too-many-arguments, t
             # url- and base64-encode the metadata.
             params["metadata"] = url_encode_dict(metadata, name="metadata", size_limit_bytes=1024)
 
+        params["_request_timeout"] = DEFAULT_REQUEST_TIMEOUT
+
         # If no inspection_id is provided, we submit the image query using image_queries_api (autogenerated via OpenAPI)
         # However, our autogenerated code does not currently support inspection_id, so if an inspection_id was
         # provided, we use the private API client instead.

src/groundlight/internalapi.py

@@ -1,3 +1,4 @@
+from datetime import datetime
 import json
 import logging
 import os
@@ -128,6 +129,8 @@ def decorated(*args, **kwargs):  # pylint: disable=inconsistent-return-statement
 
             while retry_count <= self.max_retries:
                 try:
+                    friendly_datestamp = datetime.now().isoformat()
+                    print(f"{friendly_datestamp} Trying {function.__name__} with args={args} kwargs={kwargs}")
                     return function(*args, **kwargs)
                 except ApiException as e:
                     is_retryable = (e.status is not None) and (e.status in self.status_code_range)
@@ -139,15 +142,16 @@ def decorated(*args, **kwargs):  # pylint: disable=inconsistent-return-statement
                     if is_retryable:
                         status_code = e.status
                         if status_code in self.status_code_range:
+                            # This is implementing a full jitter strategy
+                            random_delay = random.uniform(0, delay)
                             logger.warning(
                                 (
                                     f"Current HTTP response status: {status_code}. "
-                                    f"Remaining retries: {self.max_retries - retry_count}"
+                                    f"Remaining retries: {self.max_retries - retry_count}. "
+                                    f"Delaying {random_delay:.1f}s before retrying."
                                 ),
                                 exc_info=True,
                             )
-                            # This is implementing a full jitter strategy
-                            random_delay = random.uniform(0, delay)
                             time.sleep(random_delay)
 
                 retry_count += 1