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

1 Skipped Deployment

Project	Deployment	Review	Updated (UTC)
docs	Skipped		Dec 20, 2025 9:41pm

Greptile Summary

This PR adds comprehensive audit logging for all billing operations by creating a new usage_log table that tracks individual billing line items across workflows, wand, and copilot usage.

Key Changes:

• Created usage_log table with proper indexes for efficient querying by user, timestamp, source, workspace, and workflow
• Implemented three logging functions: logModelUsage() for token-based charges, logFixedUsage() for flat fees, and logWorkflowUsageBatch() for performance-optimized batch inserts
• Integrated audit logging into all billing touchpoints: workflow executions, wand operations, and copilot API calls
• Added new GET endpoint /api/users/me/usage-logs with filtering and cursor-based pagination
• Cursor pagination correctly handles duplicate timestamps using compound ordering on (createdAt, id)
• Error handling ensures audit logging failures don't break main billing flow

Implementation Quality:

• Uses batch inserts for workflow executions to minimize database overhead
• Proper foreign key constraints with SET NULL on deletion for optional references
• Non-blocking error handling - audit failures are logged but don't throw
• Comprehensive indexes support common query patterns

Confidence Score: 5/5

• This PR is safe to merge with minimal risk
• The implementation is well-architected with proper error handling, database indexes, and pagination logic. The previous cursor pagination issue has been fixed with compound ordering. All logging functions are non-blocking and won't break billing operations if they fail. The schema design follows best practices with appropriate foreign key constraints and indexes.
• No files require special attention

Important Files Changed

Sequence Diagram

sequenceDiagram
    participant Client
    participant API as API Endpoint
    participant Logger as Usage Logger
    participant UserStats as User Stats Table
    participant UsageLog as Usage Log Table
    participant Billing as Billing Service

    Note over Client,Billing: Workflow Execution Flow
    Client->>API: Execute Workflow
    API->>Logger: ExecutionLogger.updateUserStats()
    Logger->>UserStats: Update aggregate stats (cost, tokens)
    Logger->>Logger: logWorkflowUsageBatch()
    Logger->>UsageLog: Batch insert (execution fee + per-model entries)
    Logger->>Billing: checkAndBillOverageThreshold()
    API-->>Client: Execution complete

    Note over Client,Billing: Wand Usage Flow
    Client->>API: POST /api/wand
    API->>UserStats: Update aggregate wand stats
    API->>Logger: logModelUsage(source='wand')
    Logger->>UsageLog: Insert single model usage entry
    API->>Billing: checkAndBillOverageThreshold()
    API-->>Client: Wand response

    Note over Client,Billing: Copilot Usage Flow
    Client->>API: POST /api/billing/update-cost
    API->>UserStats: Update aggregate copilot stats
    API->>Logger: logModelUsage(source='copilot')
    Logger->>UsageLog: Insert single model usage entry
    API->>Billing: checkAndBillOverageThreshold()
    API-->>Client: Cost update confirmed

    Note over Client,Billing: Query Usage Logs Flow
    Client->>API: GET /api/users/me/usage-logs
    API->>Logger: getUserUsageLogs()
    Logger->>UsageLog: Query with filters (source, workspace, date range)
    UsageLog-->>Logger: Paginated logs + summary
    Logger-->>API: Usage logs result
    API-->>Client: Return logs with pagination

@greptile

@greptile