fix(core): Ensure zero-quota limits fail fast to prevent retry loop hang

[luisfelipe-alt]

Fail fast on zero-quota limits to prevent retry loop hang

Summary

This PR fixes a critical bug in the error classification logic where the Gemini-CLI gets trapped in a futile 10-attempt retry loop when hitting a hard quota limit of 0 (e.g., on unbilled free-tier accounts). This bug causes the CLI to hang for up to 10 minutes before failing, severely degrading the user experience.

Details

• The Bug: When the Gemini API returns a 429 error with limit: 0 and a "Please retry in 59.906331105s." message, classifyGoogleError matches the retry delay and incorrectly classifies it as a RetryableQuotaError.
• The Fix: We added a defensive check in classifyGoogleError (inside packages/core/src/utils/googleQuotaErrors.ts) to intercept any error message containing "limit: 0" or "limit:0" and immediately classify it as a TerminalQuotaError. This check is executed before the retry delay regex is evaluated.
• Impact: The CLI now fails fast on zero-quota limits, immediately propagating the error to the user instead of hanging for 10 minutes. Standard rate limits (which do not contain limit: 0) are unaffected and will still be retried with backoff as expected.

Related Issues

How to Validate

1. Run the newly added unit test case to verify that limit: 0 errors are correctly classified as TerminalQuotaError:

   npx vitest run packages/core/src/utils/googleQuotaErrors.test.ts

   Expected Result: All 35 tests pass successfully, including the new test case should return TerminalQuotaError when limit is 0 even if message contains "Please retry in Xs".

2. Run the retry utility tests to ensure no regressions in the core retry logic:

   npx vitest run packages/core/src/utils/retry.test.ts

   Expected Result: All 41 tests pass successfully.

Pre-Merge Checklist

• Updated relevant documentation and README (if needed)
• Added/updated tests (if needed)
• Noted breaking changes (if any)
• Validated on required platforms/methods:
  • MacOS
    • npm run
    • npx
    • Docker
    • Podman
    • Seatbelt
  • Windows
    • npm run
    • npx
    • Docker
  • Linux
    • npm run
    • npx
    • Docker

[gemini-code-assist]

Summary of Changes

Hello, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request addresses a performance issue where the Gemini-CLI would incorrectly classify hard quota limits (limit: 0) as retryable errors. By intercepting these specific error messages before the retry logic evaluates them, the CLI now terminates immediately, significantly improving the user experience by avoiding unnecessary and futile retry attempts.

Highlights

• Error Classification Logic: Added a defensive check in classifyGoogleError to identify 'limit: 0' quota errors and immediately classify them as TerminalQuotaError.
• Retry Loop Prevention: Prevents the CLI from entering a 10-minute retry loop when encountering hard quota limits by failing fast instead of attempting retries.
• Testing: Added a new unit test case to verify that errors containing 'limit: 0' are correctly handled even when they include retry delay messages.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize the Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counterproductive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for GitHub and other Google products, sign up here.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[github-actions]

📊 PR Size: size/M

• Lines changed: 157
• Additions: +148
• Deletions: -9
• Files changed: 2

[gemini-code-assist]

Code Review

This pull request introduces a check to classify Google API errors with a limit of 0 as terminal quota errors, even if they contain a retry message, and adds a corresponding unit test. The feedback suggests replacing the simple string inclusion check with a regular expression to avoid false positives on fractional or padded limits (e.g., 'limit: 0.5').

[luisfelipe-alt]

/gemini review

[gemini-code-assist]

Code Review

This pull request introduces a check to classify Google API errors with a limit of 0 as a TerminalQuotaError, preventing infinite retry loops when a quota is completely exhausted. It also adds a unit test to verify this behavior. The reviewer identified a critical issue where the check is placed inside a fallback block, meaning structured errors with details will bypass it and still be incorrectly classified as retryable. Additionally, the reviewer pointed out that the new test masks this issue by mocking the error parser to return null, and suggested moving the check earlier in the logic and adding a test case with structured RetryInfo details.

[luisfelipe-alt]

/gemini review

[gemini-code-assist]

Code Review

This pull request refactors the Google quota error classification logic in packages/core/src/utils/googleQuotaErrors.ts by extracting the errorMessage variable to the top of the function and moving the universal limit: 0 check before the fallback block. It also adds corresponding unit tests in packages/core/src/utils/googleQuotaErrors.test.ts to ensure TerminalQuotaError is correctly returned when the limit is 0, even if retry information is present. There are no review comments provided, and I have no feedback to provide.

[galdawave]

Code Review: PR #27698

Intent Summary: This PR introduces a fail-fast mechanism for Gemini API errors that indicate a zero-quota limit (limit: 0). By intercepting this specific error signature early, it correctly classifies it as a TerminalQuotaError and prevents the CLI from entering a futile, long-running retry loop.

🚨 Critical Concerns (P0/P1)

Action required before merging.

• packages/core/src/utils/googleQuotaErrors.ts:244: The condition status === undefined paired with the loose regex /limit:\s*0(?!\d|\.)/ is dangerously greedy. Because classifyGoogleError takes error: unknown, any unrelated application error containing the string "limit: 0" (e.g., new Error("Validation failed: array length limit: 0")) will lack a status and be incorrectly swallowed and misclassified as a TerminalQuotaError (a Google API quota issue).

    if (
      (status === 429 ||
        status === 499 ||
        status === 503) &&
      /limit:\s*0(?!\d|\.)/.test(errorMessage)
    ) {
  

  Note: If you absolutely must catch zero-quota errors that lack an HTTP status, the regex must be significantly tightened to avoid false positives (e.g., `/Quota exceeded.limit:\s0/).

🧹 Refactoring & Nits (P2/P3)

Recommended improvements.

• packages/core/src/utils/googleQuotaErrors.test.ts:832: AI Slop. Passing an empty new Error() and relying purely on the mock makes the test non-intuitive. Pass a representative error to make the test self-documenting.

      const result = classifyGoogleError(new Error('Quota exceeded for limit: 0'));
  

• packages/core/src/utils/googleQuotaErrors.ts:238: Minor DRY violation. If you move the errorMessage variable definition to the very top of the function (before the status === 404 check), you can reuse it for the 404 error fallback (errorMessage || 'Model not found'), removing duplicated message extraction logic.

📝 Metadata Review

Feedback on PR description or commit message clarity.

• The PR description is excellent. It clearly defines the bug, the exact fix, the impact, and provides explicit instructions for test validation.

[luisfelipe-alt]

Code Review: PR #27698

Intent Summary: This PR introduces a fail-fast mechanism for Gemini API errors that indicate a zero-quota limit (limit: 0). By intercepting this specific error signature early, it correctly classifies it as a TerminalQuotaError and prevents the CLI from entering a futile, long-running retry loop.

🚨 Critical Concerns (P0/P1)

Action required before merging.

• packages/core/src/utils/googleQuotaErrors.ts:244: The condition status === undefined paired with the loose regex /limit:\s*0(?!\d|\.)/ is dangerously greedy. Because classifyGoogleError takes error: unknown, any unrelated application error containing the string "limit: 0" (e.g., new Error("Validation failed: array length limit: 0")) will lack a status and be incorrectly swallowed and misclassified as a TerminalQuotaError (a Google API quota issue).

    if (
      (status === 429 ||
        status === 499 ||
        status === 503) &&
      /limit:\s*0(?!\d|\.)/.test(errorMessage)
    ) {
  

  _Note: If you absolutely must catch zero-quota errors that lack an HTTP status, the regex must be significantly tightened to avoid false positives (e.g., `/Quota exceeded.limit:\s_0/).

🧹 Refactoring & Nits (P2/P3)

Recommended improvements.

• packages/core/src/utils/googleQuotaErrors.test.ts:832: AI Slop. Passing an empty new Error() and relying purely on the mock makes the test non-intuitive. Pass a representative error to make the test self-documenting.

      const result = classifyGoogleError(new Error('Quota exceeded for limit: 0'));
  

• packages/core/src/utils/googleQuotaErrors.ts:238: Minor DRY violation. If you move the errorMessage variable definition to the very top of the function (before the status === 404 check), you can reuse it for the 404 error fallback (errorMessage || 'Model not found'), removing duplicated message extraction logic.

📝 Metadata Review

Feedback on PR description or commit message clarity.

• The PR description is excellent. It clearly defines the bug, the exact fix, the impact, and provides explicit instructions for test validation.

Done, I have attended the comments:

1. Tightened Regex: We tightened the regular expression to /(?:Quota exceeded|quota_exceeded).limit:\s0(?!\d|.)/i. This ensures that even if status is undefined, the error must contain both "Quota exceeded" (or "quota_exceeded") and "limit: 0" to be classified as a quota error, completely avoiding false positives on unrelated application errors (like "Validation failed: array length limit: 0").

2. DRY Refactoring: We moved the errorMessage extraction to the very top of classifyGoogleError, allowing us to reuse it in the status === 404 check and removing duplicated logic.

3. Self-Documenting Test Case: We updated the test case to pass a representative error message (new Error('Quota exceeded for limit: 0')) to classifyGoogleError, making the test self-documenting and intuitive.

[luisfelipe-alt]

/gemini review

[gemini-code-assist]

Code Review

This pull request refactors classifyGoogleError to extract a common errorMessage variable and introduces a universal check to return a TerminalQuotaError when a quota limit of 0 is detected, accompanied by new unit tests. The reviewer identified a high-severity issue in the regular expression used for this check, noting that .* does not match newline characters in JavaScript, which could lead to failures in production. They provided a code suggestion to use [\s\S]* instead for robust matching.

[luisfelipe-alt]

/gemini review

[gemini-code-assist]

Code Review

This pull request refactors Google quota error classification by extracting the error message extraction logic and introducing a universal check for "limit: 0" to classify errors as TerminalQuotaError, supported by new unit tests. However, a critical syntax error was introduced in the newly added if statement with a duplicate closing parenthesis and opening brace, which will cause compilation to fail.

[shaman2527]

si me pasa igual cuando entro con la cuenta en gemini cli hay que esperar como 10 min para que funcione

[luisfelipe-alt]

/gemini review

[gemini-code-assist]

Code Review

This pull request introduces a universal check for 'limit: 0' errors in classifyGoogleError to return a TerminalQuotaError early, along with corresponding unit tests. The review feedback highlights two critical issues: first, moving the errorMessage definition to the top of the function causes a regression in 404 error handling where plain object errors evaluate to '[object Object]' instead of falling back to 'Model not found'; second, the negative lookahead in the regex incorrectly rejects matches when 'limit: 0' is followed by a period at the end of a sentence, which can be resolved by updating the lookahead pattern to allow periods not followed by a digit.

[luisfelipe-alt]

/gemini review

[luisfelipe-alt]

/gemini review

[gemini-code-assist]

Code Review

This pull request introduces a universal check for limit: 0 to classify certain errors as TerminalQuotaError and refactors classifyGoogleError to extract the error message earlier, supported by several new test cases. The review feedback identifies a critical issue where plain object errors (which are not instances of Error) will have their messages evaluated as "[object Object]", bypassing the limit: 0 check and obscuring error messages. A robust extraction of the message property from plain objects is suggested to resolve this.

[luisfelipe-alt]

/gemini review

[gemini-code-assist]

Code Review

This pull request refactors the Google error classification logic in packages/core/src/utils/googleQuotaErrors.ts by centralizing the extraction of errorMessage and introducing a universal check to classify limit: 0 errors (for status codes 429, 499, and 503) as TerminalQuotaError. It also adds comprehensive unit tests in googleQuotaErrors.test.ts to validate this new classification behavior under various conditions. There are no review comments, and I have no feedback to provide.

[luisfelipe-alt]

/gemini review

[gemini-code-assist]

Code Review

This pull request refactors the Google error classification logic in packages/core/src/utils/googleQuotaErrors.ts by introducing a helper function extractErrorMessage to consistently extract error messages from various error formats. It also implements a universal check to classify quota errors with a limit of 0 as TerminalQuotaError across status codes 429, 499, and 503, ensuring fractional limits (like 0.5) are still treated as retryable. Additionally, comprehensive unit tests have been added to validate these changes under different error structures and edge cases. No review comments were provided, so there is no further feedback to address.