Upload Analytics

The upload analytics dashboard tracks document upload activity across your workspace. It shows how many documents have been uploaded, their processing status, failure rates, and common issues. This helps administrators identify problems with the document pipeline and ensure that the upload system is working smoothly. Admin role is required.

Accessing Upload Analytics

Navigate to /admin/upload-analytics from the admin sidebar.

Upload analytics dashboard with processing metrics and status breakdown

Key Metrics

The dashboard presents several top-level metrics:

• Total Uploads -- The cumulative number of documents uploaded to your workspace.
• Uploads This Period -- Documents uploaded in the current billing period (month), with a comparison to the previous period.
• Processing Success Rate -- The percentage of uploaded documents that were successfully parsed, chunked, and indexed. A healthy system maintains a success rate above 95%.
• Average Processing Time -- The mean time from upload to full indexing (text extraction, chunking, embedding generation).
• Pending -- Documents currently in the processing queue awaiting text extraction or embedding generation.

Upload Activity Chart

A timeline chart shows daily upload volume over the selected date range. Use this to identify:

• Peak upload periods (e.g., at the start of semesters or near grant deadlines).
• Unusual drops in upload activity that may indicate system issues.
• Correlation between upload volume and processing delays.

File Type Breakdown

A breakdown chart shows uploads by file type:

• PDF (typically the majority of academic document uploads).
• Word documents (DOCX).
• Plain text (TXT).
• Rich text format (RTF).
• Other supported formats.

This data helps you understand your workspace's document composition and anticipate format-specific processing issues.

Processing Status Details

Each document goes through several processing stages:

1. Uploaded -- The file is received and stored.
2. Text Extracted -- The document text has been successfully parsed.
3. Chunked -- The text has been split into semantic chunks.
4. Embedded -- Vector embeddings have been generated for all chunks.
5. Indexed -- The document is fully searchable and available for AI features.

The processing status section shows how many documents are at each stage and highlights any that are stuck or failed.

Failed Uploads

The failed uploads table lists documents that could not be processed, with:

• Document name and uploader.
• Failure stage (which processing step failed).
• Error message.
• Timestamp.

Common failure reasons include:

• Corrupted PDF files that cannot be parsed.
• Password-protected documents.
• Unsupported file formats.
• Documents that exceed the maximum file size.

Tips

• Check the failed uploads list weekly and notify users if their documents need to be re-uploaded in a different format.
• If the processing success rate drops below 95%, investigate whether a specific file type or processing stage is causing failures.
• Monitor the processing queue during high-upload periods to ensure documents are being indexed in a timely manner.
• Use the file type breakdown to identify whether your workspace needs support for additional document formats.