openclaw - ✅(Solved) Fix fix(agents): provider-specific errors misclassified during failover [2 pull requests, 1 participants]

openclaw2026-04-01 07:51:13

ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

GitHub issue URL

Helpful · Quick feedback

GitHub stats

openclaw/openclaw#58837•Fetched 2026-04-08 02:32:03

View on GitHub

Comments

Participants

Timeline

Reactions

Author

aaron-he-zhu

Participants

aaron-he-zhu

Timeline (top)

cross-referenced ×2closed ×1locked ×1

Error Message

classifyFailoverReason() in src/agents/pi-embedded-helpers/errors.ts relies on generic string matching that misclassifies errors from several providers. This causes the failover engine to choose wrong recovery strategies — for example, retrying with the same oversized context instead of triggering compaction, or treating a rate limit as an unknown error. | Provider | Error Message | Expected | Actual | Provider-specific error patterns should be normalized to the existing FailoverReason enum so the failover engine can make correct decisions regardless of which provider returned the error.

Fix Action

Fixed

Fixed by PR: fix(agents): normalize provider errors for better failover (https://github.com/openclaw/openclaw/pull/58845)
Fixed by PR: fix(agents): normalize provider errors for better failover (https://github.com/openclaw/openclaw/pull/58856)

PR fix notes

PR #58845: fix(agents): normalize provider errors for better failover

Repository: openclaw/openclaw
Author: aaron-he-zhu
State: closed | merged: False
Link: https://github.com/openclaw/openclaw/pull/58845

Description (problem / solution / changelog)

Fixes #58837

Related: #58347 #58305 #58377 #58392 (LiveSessionModelSwitchError cluster)

Summary

Add provider-specific error patterns for AWS Bedrock, Ollama, Mistral, Cohere, DeepSeek, Together AI, and Cloudflare Workers AI. These providers return errors in non-standard formats that the generic classifiers miss, causing incorrect failover behavior.

Create provider-error-patterns.ts with provider-specific matchers
matchesProviderContextOverflow() catches context overflow patterns the generic regex misses
classifyProviderSpecificError() maps provider errors to FailoverReason
Wire into isContextOverflowError() and classifyFailoverReason() as catch-all layers

Examples fixed:

Provider	Error	Was	Now
Bedrock	`ThrottlingException`	unclassified	`rate_limit`
Bedrock	`ModelNotReadyException`	unclassified	`overloaded`
Groq	`model_is_deactivated`	unclassified	`model_not_found`
Together AI	`concurrency limit reached`	unclassified	`rate_limit`

Test plan

Unit tests for matchesProviderContextOverflow (8 provider samples + negatives)
Unit tests for classifyProviderSpecificError (5 provider mappings)
Integration tests with isContextOverflowError and classifyFailoverReason
Existing failover tests pass (35 tests)

Changed files

src/agents/pi-embedded-helpers/errors.ts (modified, +12/-1)
src/agents/pi-embedded-helpers/provider-error-patterns.test.ts (added, +92/-0)
src/agents/pi-embedded-helpers/provider-error-patterns.ts (added, +117/-0)

PR #58856: fix(agents): normalize provider errors for better failover

Repository: openclaw/openclaw
Author: aaron-he-zhu
State: closed | merged: True
Link: https://github.com/openclaw/openclaw/pull/58856

Description (problem / solution / changelog)

Fixes #58837

Related: #58347 #58305 #58377 #58392 (LiveSessionModelSwitchError cluster)

Summary

Create provider-error-patterns.ts with provider-specific matchers
matchesProviderContextOverflow() catches context overflow patterns the generic regex misses
classifyProviderSpecificError() maps provider errors to FailoverReason
Wire into isContextOverflowError() and classifyFailoverReason() as catch-all layers

Examples fixed:

Provider	Error	Was	Now
Bedrock	`ThrottlingException`	unclassified	`rate_limit`
Bedrock	`ModelNotReadyException`	unclassified	`overloaded`
Groq	`model_is_deactivated`	unclassified	`model_not_found`
Together AI	`concurrency limit reached`	unclassified	`rate_limit`

Test plan

Unit tests for matchesProviderContextOverflow (8 provider samples + negatives)
Unit tests for classifyProviderSpecificError (5 provider mappings)
Integration tests with isContextOverflowError and classifyFailoverReason
Existing failover tests pass (35 tests)
tsgo --noEmit clean, oxlint clean, oxfmt --check clean

Changed files

CHANGELOG.md (modified, +1/-0)
src/agents/failover-error.test.ts (modified, +18/-0)
src/agents/pi-embedded-helpers.isbillingerrormessage.test.ts (modified, +29/-0)
src/agents/pi-embedded-helpers/errors.ts (modified, +96/-79)
src/agents/pi-embedded-helpers/provider-error-patterns.test.ts (added, +100/-0)
src/agents/pi-embedded-helpers/provider-error-patterns.ts (added, +111/-0)

RAW_BUFFERClick to expand / collapse

Bug Description

Examples of Misclassification

Provider	Error Message	Expected	Actual
AWS Bedrock	`ThrottlingException: Too many concurrent requests`	`rate_limit`	`null` (unclassified)
AWS Bedrock	`ModelNotReadyException: model is not ready`	`overloaded`	`null`
Groq	`model_is_deactivated`	`model_not_found`	`null`
Together AI	`concurrency limit has been reached`	`rate_limit`	`null`
Google Vertex	`INVALID_ARGUMENT: exceeds the maximum input tokens`	context overflow	`null`

Impact

Incorrect failover decisions waste API credits (retrying instead of switching model/provider)
May contribute to LiveSessionModelSwitchError issues (#58347, #58305, #58377, #58392) when failover misclassifies errors and picks wrong recovery paths
Users on providers like Bedrock, Together AI, Groq see unexpected failures instead of graceful failover

Expected Behavior

Provider-specific error patterns should be normalized to the existing FailoverReason enum so the failover engine can make correct decisions regardless of which provider returned the error.

Affected Code

src/agents/pi-embedded-helpers/errors.ts:1002 — classifyFailoverReason()
src/agents/pi-embedded-helpers/failover-matches.ts — generic pattern matchers

extent analysis

TL;DR

Update the classifyFailoverReason() function in src/agents/pi-embedded-helpers/errors.ts to include provider-specific error patterns.

Guidance

Review the classifyFailoverReason() function to identify areas where generic string matching is used and replace it with provider-specific error patterns.
Update the failover-matches.ts file to include additional pattern matchers for each provider's unique error messages.
Verify the changes by testing the failover engine with various error scenarios from different providers.
Consider adding unit tests to ensure the classifyFailoverReason() function correctly classifies errors for each provider.

Example

// Example of updated classifyFailoverReason() function
function classifyFailoverReason(errorMessage: string): FailoverReason {
  // Provider-specific error patterns
  if (errorMessage.includes('ThrottlingException: Too many concurrent requests')) {
    return FailoverReason.rate_limit;
  } else if (errorMessage.includes('ModelNotReadyException: model is not ready')) {
    return FailoverReason.overloaded;
  } else if (errorMessage.includes('model_is_deactivated')) {
    return FailoverReason.model_not_found;
  }
  // ...
}

Notes

The updated classifyFailoverReason() function should handle errors from various providers, including AWS Bedrock, Groq, Together AI, and Google Vertex.

Recommendation

Apply workaround: Update the classifyFailoverReason() function to include provider-specific error patterns, as this will allow the failover engine to make correct decisions regardless of which provider returned the error.

Vote matrix · Quick signals

Works

Did the solution work? Tap to confirm.

Easy Fix

Was it a quick fix?

Time Saver

Did it save you time?

Blocking

Was it severely blocking?

Common Issue

Are others likely hitting this too?

Flaky / Intermittent

Is it intermittent?

Verified / Reproducible

Can you reproduce it reliably?

#api #serialization error #model compatibility #GPU setup #container setup

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

openclaw - ✅(Solved) Fix fix(agents): provider-specific errors misclassified during failover [2 pull requests, 1 participants]

Recommended Tools

GitHub issue graph ai analysis

Error Message

Fix Action

Fixed

PR fix notes

PR #58845: fix(agents): normalize provider errors for better failover

Description (problem / solution / changelog)

Summary

Test plan

Changed files

PR #58856: fix(agents): normalize provider errors for better failover

Description (problem / solution / changelog)

Summary

Test plan

Changed files

Bug Description

Examples of Misclassification

Impact

Expected Behavior

Affected Code

extent analysis

TL;DR

Guidance

Example

Notes

Recommendation

Still need to ship something?

RELATED_DISCOVERY

TRENDING