diff --git a/litellm/__init__.py b/litellm/__init__.py
index 91457f9b049..84fc7fc6789 100644
--- a/litellm/__init__.py
+++ b/litellm/__init__.py
@@ -7,6 +7,7 @@ import threading
 import os
 from typing import Callable, List, Optional, Dict, Union, Any, Literal, get_args
 from litellm.llms.custom_httpx.http_handler import AsyncHTTPHandler, HTTPHandler
+from litellm import constants
 from litellm.caching.caching import Cache, DualCache, RedisCache, InMemoryCache
 from litellm.types.llms.bedrock import COHERE_EMBEDDING_INPUT_TYPES
 from litellm.types.utils import (
diff --git a/litellm/proxy/proxy_server.py b/litellm/proxy/proxy_server.py
index 421a588e390..015292db183 100644
--- a/litellm/proxy/proxy_server.py
+++ b/litellm/proxy/proxy_server.py
@@ -1972,6 +1972,13 @@ class ProxyConfig:
                         raise Exception(
                             f"Invalid value set for upperbound_key_generate_params - value={value}"
                         )
+                elif key == "constants":
+                    # allow overriding constants
+                    for constant_key, constant_value in value.items():
+                        verbose_proxy_logger.debug(
+                            f"{blue_color_code} setting litellm.constants.{constant_key}={constant_value}{reset_color_code}"
+                        )
+                        setattr(litellm.constants, constant_key, constant_value)
                 else:
                     verbose_proxy_logger.debug(
                         f"{blue_color_code} setting litellm.{key}={value}{reset_color_code}"
@@ -8850,3 +8857,9 @@ app.include_router(openai_files_router)
 app.include_router(team_callback_router)
 app.include_router(budget_management_router)
 app.include_router(model_management_router)
+
+app.include_router(ui_crud_endpoints_router)
+app.include_router(openai_files_router)
+app.include_router(team_callback_router)
+app.include_router(budget_management_router)
+app.include_router(model_management_router)
diff --git a/litellm/router.py b/litellm/router.py
index 722eed56f1e..0dd8e2ec928 100644
--- a/litellm/router.py
+++ b/litellm/router.py
@@ -68,7 +68,6 @@ from litellm.router_utils.batch_utils import (
 from litellm.router_utils.client_initalization_utils import InitalizeOpenAISDKClient
 from litellm.router_utils.cooldown_cache import CooldownCache
 from litellm.router_utils.cooldown_handlers import (
-    DEFAULT_COOLDOWN_TIME_SECONDS,
     _async_get_cooldown_deployments,
     _async_get_cooldown_deployments_with_debug_info,
     _get_cooldown_deployments,
@@ -383,7 +382,9 @@ class Router:
             self.allowed_fails = allowed_fails
         else:
             self.allowed_fails = litellm.allowed_fails
-        self.cooldown_time = cooldown_time or DEFAULT_COOLDOWN_TIME_SECONDS
+        self.cooldown_time = (
+            cooldown_time or litellm.constants.DEFAULT_COOLDOWN_TIME_SECONDS
+        )
         self.cooldown_cache = CooldownCache(
             cache=self.cache, default_cooldown_time=self.cooldown_time
         )
diff --git a/litellm/router_utils/cooldown_handlers.py b/litellm/router_utils/cooldown_handlers.py
index 8f5c3895a6d..4cbb3895af3 100644
--- a/litellm/router_utils/cooldown_handlers.py
+++ b/litellm/router_utils/cooldown_handlers.py
@@ -11,11 +11,6 @@ from typing import TYPE_CHECKING, Any, List, Optional, Union
 
 import litellm
 from litellm._logging import verbose_router_logger
-from litellm.constants import (
-    DEFAULT_COOLDOWN_TIME_SECONDS,
-    DEFAULT_FAILURE_THRESHOLD_PERCENT,
-    SINGLE_DEPLOYMENT_TRAFFIC_FAILURE_THRESHOLD,
-)
 from litellm.router_utils.cooldown_callbacks import router_cooldown_event_callback
 
 from .router_callbacks.track_deployment_metrics import (
@@ -197,12 +192,12 @@ def _should_cooldown_deployment(
         elif (
             percent_fails == 1.0
             and total_requests_this_minute
-            >= SINGLE_DEPLOYMENT_TRAFFIC_FAILURE_THRESHOLD
+            >= litellm.constants.SINGLE_DEPLOYMENT_TRAFFIC_FAILURE_THRESHOLD
         ):
             # Cooldown if all requests failed and we have reasonable traffic
             return True
         elif (
-            percent_fails > DEFAULT_FAILURE_THRESHOLD_PERCENT
+            percent_fails > litellm.constants.DEFAULT_FAILURE_THRESHOLD_PERCENT
             and not is_single_deployment_model_group  # by default we should avoid cooldowns on single deployment model groups
         ):
             return True
@@ -377,7 +372,8 @@ def should_cooldown_based_on_allowed_fails_policy(
         or litellm_router_instance.allowed_fails
     )
     cooldown_time = (
-        litellm_router_instance.cooldown_time or DEFAULT_COOLDOWN_TIME_SECONDS
+        litellm_router_instance.cooldown_time
+        or litellm.constants.DEFAULT_COOLDOWN_TIME_SECONDS
     )
 
     current_fails = litellm_router_instance.failed_calls.get_cache(key=deployment) or 0