LMs: retry with exponential backoff for a limited set of error codes #1753
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Signed-off-by: dbczumar <[email protected]>
Signed-off-by: dbczumar <[email protected]>
Signed-off-by: dbczumar <[email protected]>
Signed-off-by: dbczumar <[email protected]>
Signed-off-by: dbczumar <[email protected]>
dbczumar
commented
Nov 5, 2024
| @@ -32,7 +33,7 @@ def __init__( | |||
| cache: bool = True, | |||
| launch_kwargs: Optional[Dict[str, Any]] = None, | |||
| callbacks: Optional[List[BaseCallback]] = None, | |||
| num_retries: int = 3, | |||
There was a problem hiding this comment.
3 retries (equivalent to < 5 seconds) is insufficient to overcome rate limiting in many production environments with high traffic
dbczumar
commented
Nov 5, 2024
| @@ -32,7 +33,7 @@ def __init__( | |||
| cache: bool = True, | |||
| launch_kwargs: Optional[Dict[str, Any]] = None, | |||
| callbacks: Optional[List[BaseCallback]] = None, | |||
| num_retries: int = 3, | |||
| num_retries: int = 8, | |||
There was a problem hiding this comment.
Empirically, 8 retries translates to ~ 1 minute of wall clock time, which should be sufficient to overcome rate limiting in most cases
Signed-off-by: dbczumar <[email protected]>
dbczumar
commented
Nov 5, 2024
Comment on lines
+199
to
+224
| retry_policy = RetryPolicy( | ||
| TimeoutErrorRetries=num_retries, | ||
| RateLimitErrorRetries=num_retries, | ||
| InternalServerErrorRetries=num_retries, | ||
| # We don't retry on errors that are unlikely to be transient | ||
| # (e.g. bad request, invalid auth credentials) | ||
| BadRequestErrorRetries=0, | ||
| AuthenticationErrorRetries=0, | ||
| ContentPolicyViolationErrorRetries=0, | ||
| ) | ||
|
|
||
| return Router( | ||
| # LiteLLM routers must specify a `model_list`, which maps model names passed | ||
| # to `completions()` into actual LiteLLM model names. For our purposes, the | ||
| # model name is the same as the LiteLLM model name, so we add a single | ||
| # entry to the `model_list` that maps the model name to itself | ||
| model_list=[ | ||
| { | ||
| "model_name": model, | ||
| "litellm_params": { | ||
| "model": model, | ||
| }, | ||
| } | ||
| ], | ||
| retry_policy=retry_policy, | ||
| ) |
There was a problem hiding this comment.
LiteLLM Routers appear to be the only mechanism allowing exponential backoff and configurable retry codes. Docs: https://docs.litellm.ai/docs/routing
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
LMs: retry with exponential backoff for a limited set of error codes