shumengya/sproutclaw
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix(ai): disable hidden provider 429 retries (#4991)

Browse Source
This commit is contained in:
Armin Ronacher
2026-05-26 23:41:58 +02:00
committed by GitHub
parent 96f0edd02b
commit 8fb1e877cd
10 changed files with 161 additions and 31 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
packages/coding-agent/docs/settings.md
View File

@@ -101,11 +101,13 @@ Set `PI_SKIP_VERSION_CHECK=1` to disable the Pi version update check. Use `--off
| `retry.maxRetries` | number | `3` | Maximum agent-level retry attempts |
| `retry.baseDelayMs` | number | `2000` | Base delay for agent-level exponential backoff (2s, 4s, 8s) |
| `retry.provider.timeoutMs` | number | SDK default | Provider/SDK request timeout in milliseconds |
| `retry.provider.maxRetries` | number | SDK default | Provider/SDK retry attempts |
| `retry.provider.maxRetries` | number | `0` | Provider/SDK retry attempts |
| `retry.provider.maxRetryDelayMs` | number | `60000` | Max server-requested delay before failing (60s) |
When a provider requests a retry delay longer than `retry.provider.maxRetryDelayMs` (e.g., Google's "quota will reset after 5h"), the request fails immediately with an informative error instead of waiting silently. Set to `0` to disable the cap.
Keep `retry.provider.maxRetries` at `0` unless provider-level retries are explicitly needed. Setting it above `0` can make SDK/provider retries handle out-of-usage-limit errors before Pi sees them, which may block the agent until the provider quota resets in some circumstances.
```json
{
"retry": {

7
packages/coding-agent/src/core/agent-session.ts
View File

@@ -2422,6 +2422,12 @@ export class AgentSession {
// Auto-Retry
// =========================================================================
private _isNonRetryableProviderLimitError(errorMessage: string): boolean {
return /GoUsageLimitError|FreeUsageLimitError|Monthly usage limit reached|available balance|insufficient_quota|out of budget|quota exceeded|billing/i.test(
errorMessage,
);
}
/**
* Check if an error is retryable (overloaded, rate limit, server errors).
* Context overflow errors are NOT retryable (handled by compaction instead).
@@ -2434,6 +2440,7 @@ export class AgentSession {
if (isContextOverflow(message, contextWindow)) return false;
const err = message.errorMessage;
if (this._isNonRetryableProviderLimitError(err)) return false;
// Match: overloaded_error, provider returned error, rate limit, 429, 500, 502, 503, 504, service unavailable, network/connection errors (including connection lost), WebSocket transport closes/errors, fetch failed, premature stream endings, HTTP/2 closed before response, terminated, retry delay exceeded
return /overloaded|provider.?returned.?error|rate.?limit|too many requests|429|500|502|503|504|service.?unavailable|server.?error|internal.?error|network.?error|connection.?error|connection.?refused|connection.?lost|websocket.?closed|websocket.?error|other side closed|fetch failed|upstream.?connect|reset before headers|socket hang up|ended without|stream ended before message_stop|http2 request did not get a response|timed? out|timeout|terminated|retry delay/i.test(
err,