valknar/llmx
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Parse the Azure OpenAI rate limit message (#5956)

Browse Source 
Fixes #4161

Currently Codex uses a regex to parse the "Please try again in 1.898s"
OpenAI-style rate limit message, so that it can wait the correct
duration before retrying. Azure OpenAI returns a different error that
looks like "Rate limit exceeded. Try again in 35 seconds."

This PR extends the regex and parsing code to match in a more fuzzy
manner, handling anything matching the pattern "try again in
\<duration>\<unit>".
This commit is contained in:
Thomas Stokes
2025-11-02 00:33:13 +08:00
committed by GitHub
parent 91e65ac0ce
commit d9118c04bf
1 changed files with 20 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

24
codex-rs/core/src/client.rs
View File

@@ -931,8 +931,10 @@ async fn stream_from_fixture(
fn rate_limit_regex() -> &'static Regex {
static RE: OnceLock<Regex> = OnceLock::new();
// Match both OpenAI-style messages like "Please try again in 1.898s"
// and Azure OpenAI-style messages like "Try again in 35 seconds".
#[expect(clippy::unwrap_used)]
RE.get_or_init(|| Regex::new(r"Please try again in (\d+(?:\.\d+)?)(s|ms)").unwrap())
RE.get_or_init(|| Regex::new(r"(?i)try again in\s*(\d+(?:\.\d+)?)\s*(s|ms|seconds?)").unwrap())
}
fn try_parse_retry_after(err: &Error) -> Option<Duration> {
@@ -940,7 +942,8 @@ fn try_parse_retry_after(err: &Error) -> Option<Duration> {
return None;
}
// parse the Please try again in 1.898s format using regex
// parse retry hints like "try again in 1.898s" or
// "Try again in 35 seconds" using regex
let re = rate_limit_regex();
if let Some(message) = &err.message
&& let Some(captures) = re.captures(message)
@@ -950,9 +953,9 @@ fn try_parse_retry_after(err: &Error) -> Option<Duration> {
if let (Some(value), Some(unit)) = (seconds, unit) {
let value = value.as_str().parse::<f64>().ok()?;
let unit = unit.as_str();
let unit = unit.as_str().to_ascii_lowercase();
if unit == "s" {
if unit == "s" || unit.starts_with("second") {
return Some(Duration::from_secs_f64(value));
} else if unit == "ms" {
return Some(Duration::from_millis(value as u64));
@@ -1427,6 +1430,19 @@ mod tests {
assert_eq!(delay, Some(Duration::from_secs_f64(1.898)));
}
#[test]
fn test_try_parse_retry_after_azure() {
let err = Error {
r#type: None,
message: Some("Rate limit exceeded. Try again in 35 seconds.".to_string()),
code: Some("rate_limit_exceeded".to_string()),
plan_type: None,
resets_at: None,
};
let delay = try_parse_retry_after(&err);
assert_eq!(delay, Some(Duration::from_secs(35)));
}
#[test]
fn error_response_deserializes_schema_known_plan_type_and_serializes_back() {
use crate::token_data::KnownPlan;