Skip to content

LLM-Driven Extraction of Unstructured Data — Built for API Deployments & ETL Pipeline Workflows

UN-2972 [MISC] Improve global rate limit error message for better user experience (#1652)

Ritwik G057420fa4a40
diff --git a/backend/api_v2/exceptions.py b/backend/api_v2/exceptions.py
index 7dab3c178..246788bff 100644
--- a/backend/api_v2/exceptions.py
+++ b/backend/api_v2/exceptions.py
@@ -82,9 +82,7 @@ class RateLimitExceeded(APIException):
current_usage=current_usage, limit=limit
)
else:
- detail = RateLimitMessages.get_global_limit_exceeded_message(
- current_usage=current_usage, limit=limit
- )
+ detail = RateLimitMessages.get_global_limit_exceeded_message()
super().__init__(detail, code)
diff --git a/backend/api_v2/rate_limit_constants.py b/backend/api_v2/rate_limit_constants.py
index 069171cad..045ade15e 100644
--- a/backend/api_v2/rate_limit_constants.py
+++ b/backend/api_v2/rate_limit_constants.py
@@ -94,8 +94,8 @@ class RateLimitMessages:
)
GLOBAL_LIMIT_EXCEEDED_TEMPLATE = (
- "System has reached the global maximum concurrent API requests limit "
- "({current_usage}/{limit}). Please try again later."
+ "Our system is currently experiencing high load. "
+ "Please try again in a few moments."
)
LOCK_ACQUISITION_FAILED = "Failed to acquire rate limit lock. Please try again."
@@ -110,8 +110,9 @@ class RateLimitMessages:
)
@classmethod
- def get_global_limit_exceeded_message(cls, current_usage: int, limit: int) -> str:
- """Get formatted global limit exceeded message."""
- return cls.GLOBAL_LIMIT_EXCEEDED_TEMPLATE.format(
- current_usage=current_usage, limit=limit
- )
+ def get_global_limit_exceeded_message(cls) -> str:
+ """Get generic global limit exceeded message.
+
+ Returns a user-friendly message without exposing system capacity.
+ """
+ return cls.GLOBAL_LIMIT_EXCEEDED_TEMPLATE
diff --git a/docs/API_DEPLOYMENT_RATE_LIMITING.md b/docs/API_DEPLOYMENT_RATE_LIMITING.md
index a7465b530..ee5ec8c48 100644
--- a/docs/API_DEPLOYMENT_RATE_LIMITING.md
+++ b/docs/API_DEPLOYMENT_RATE_LIMITING.md
@@ -406,6 +406,7 @@ Content-Type: application/json
When rate limit is exceeded, API returns 429 with a standardized error response:
+#### Organization Limit Exceeded
```http
HTTP/1.1 429 Too Many Requests
Content-Type: application/json
@@ -422,12 +423,29 @@ Content-Type: application/json
}
```
+#### Global Limit Exceeded (System Overload)
+```http
+HTTP/1.1 429 Too Many Requests
+Content-Type: application/json
+
+{
+ "type": "client_error",
+ "errors": [
+ {
+ "code": "error",
+ "detail": "Our system is currently experiencing high load. Please try again in a few moments.",
+ "attr": null
+ }
+ ]
+}
+```
+
**Response Format**:
The error response follows the `drf-standardized-errors` format used throughout the application:
- `type`: Error type (always `"client_error"` for 429)
- `errors`: Array of error objects
- `code`: Error code (always `"error"`)
- - `detail`: Human-readable error message with current usage and limit
+ - `detail`: Human-readable error message
- `attr`: Field name (null for non-field errors)
**Note**: Clients should implement their own retry logic with exponential backoff. The rate limit will be released once active requests complete.