fix(embeddings): preserve safe provider failure diagnostics

[waleedlatif1]

Summary

• Log bounded embedding provider, model/deployment, HTTP status, recognized error codes, and request IDs internally without retaining response messages, document text, credentials, or full headers.
• Classify embedding failures, including concurrent batch aggregates, separately from unavailable source content. Preserve existing caller-facing errors, retries, quota handling, and model selection.

Type of Change

• Bug fix

Testing

• 253 tests passed across embedding requests, diagnostic sanitization, connector diagnostics, and document processing.
• Lint, application type-check, all 46 repository audits (including API boundary validation), block registry validation, and docs manifest verification passed.

Checklist

• Code follows project style guidelines
• Self-reviewed my changes
• Tests added/updated and passing
• No new warnings introduced
• I confirm that I have read and agree to the terms outlined in the Contributor License Agreement (CLA)

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

1 Skipped Deployment

Project	Deployment	Actions	Updated
docs	Skipped		Sep 17, 2026 9:32pm UTC

[greptile-apps]

Confidence Score: 5/5

The PR appears safe to merge; no actionable regression or outstanding repository-rule violation was identified.

Summary

This PR preserves bounded, sanitized diagnostics for embedding-provider failures and distinguishes wrapped embedding failures from unavailable source content.

• Moves EmbeddingAPIError into a dedicated module and adds cycle-safe traversal through causes and aggregate errors.
• Logs allowlisted provider metadata, status codes, request IDs, and machine-readable error codes without retaining free-form response content.
• Classifies wrapped embedding failures separately in connector diagnostics while preserving caller-facing errors, quota handling, and retry behavior.
• Adds coverage for provider diagnostics, malformed bodies, nested aggregates, cycles, and traversal bounds.

Diagram

%%{init: {'theme': 'neutral'}}%%
flowchart TD
  A[Embedding provider response] --> B{Response successful?}
  B -->|Yes| C[Validate and return embeddings]
  B -->|No| D[Read bounded response body]
  D --> E[Extract allowlisted code, type, and request ID]
  E --> F[Write sanitized internal diagnostic]
  F --> G[Create EmbeddingAPIError]
  G --> H{Quota exhausted?}
  H -->|Yes| I[Apply existing quota handling]
  H -->|No| J[Apply existing retry policy]
  G --> K[Cause or AggregateError wrappers]
  K --> L[Find bounded embedding failure]
  L --> M[Classify connector diagnostic as embedding]

Loading

_{Reviews (2) · Last reviewed commit: "fix(embeddings): classify aggregated bat..."}

[cubic-dev-ai]

All reported issues were addressed across 10 files

<sub>Reply with feedback, questions, or to request a fix.

Fix all with cubic | Re-trigger cubic</sub>

[waleedlatif1]

@greptile

[waleedlatif1]

@cubic-dev-ai review this PR

[cubic-dev-ai]

@cubic-dev-ai review this PR

@waleedlatif1 I have started the AI code review. It will take a few minutes to complete.

[waleedlatif1]

@cubic-dev-ai review this PR

[cubic-dev-ai]

@cubic-dev-ai review this PR

@waleedlatif1 I have started the AI code review. It will take a few minutes to complete.

[cubic-dev-ai]

No issues found across 10 files

Confidence score: 5/5

• Automated review surfaced no issues in the provided summaries.
• No files require special attention.

_{Re-trigger cubic}

[cubic-dev-ai]

No issues found across 10 files

Confidence score: 5/5

• Automated review surfaced no issues in the provided summaries.
• No files require special attention.

_{Re-trigger cubic}