Logs

/llamafarm:logs - View Service Logs

View and filter logs from LlamaFarm services for debugging and monitoring.

Usage

/llamafarm:logs                     # Recent logs from all services
/llamafarm:logs --service server    # Logs from specific service
/llamafarm:logs --follow            # Stream logs in real-time
/llamafarm:logs --errors            # Only show errors
/llamafarm:logs --since 1h          # Logs from last hour

What This Command Does

1. Retrieves logs - From server, RAG worker, and other components
2. Filters output - By service, level, time range
3. Formats for reading - Highlights errors, timestamps
4. Identifies issues - Common error patterns

Implementation

When the user runs /llamafarm:logs, follow these steps:

Step 1: Determine Log Source

LlamaFarm logs are typically in:

• Server: stdout/stderr when running in foreground
• Celery: Worker logs
• File: ~/.llamafarm/logs/ (if configured)

# Check if services are running and get log location
lf services logs --help

Step 2: Retrieve Logs

# All recent logs
lf services logs

# Specific service
lf services logs --service server
lf services logs --service rag

# Follow mode (real-time)
lf services logs --follow

Step 3: Filter and Format

For --errors flag:

lf services logs | grep -E "(ERROR|CRITICAL|Exception|Traceback)"

For --since flag:

lf services logs --since "1 hour ago"

Step 4: Present Results

LlamaFarm Logs (last 50 lines)
==============================

[2024-01-14 10:23:45] INFO  [server] Started on http://localhost:14345
[2024-01-14 10:23:46] INFO  [celery] Worker ready, consuming from queue
[2024-01-14 10:25:12] INFO  [rag] Processing dataset: research_papers
[2024-01-14 10:25:15] INFO  [rag] Parsed 5 documents, 127 chunks created
[2024-01-14 10:25:18] INFO  [rag] Embedding batch 1/3...
[2024-01-14 10:25:22] INFO  [rag] Embedding batch 2/3...
[2024-01-14 10:25:26] INFO  [rag] Embedding batch 3/3...
[2024-01-14 10:25:27] INFO  [rag] Ingestion complete: 127 chunks indexed
[2024-01-14 10:30:01] INFO  [chat] Request from session abc123
[2024-01-14 10:30:02] INFO  [rag] Query: "neural scaling laws"
[2024-01-14 10:30:02] INFO  [rag] Retrieved 10 chunks (0.15s)
[2024-01-14 10:30:05] INFO  [chat] Response generated (3.2s)

Service-Specific Logs

Server Logs

/llamafarm:logs --service server

Shows:

• HTTP requests and responses
• API errors
• Configuration loading
• Model interactions

RAG Worker Logs

/llamafarm:logs --service rag

Shows:

• Document parsing progress
• Chunking details
• Embedding generation
• Vector store operations
• Retrieval queries

Celery Logs

/llamafarm:logs --service celery

Shows:

• Task queue status
• Worker health
• Task execution times
• Queue backlogs

Error Analysis

When errors are found, provide context:

Error Found in Logs
===================

[2024-01-14 10:35:12] ERROR [rag] Failed to parse document: contract.pdf
Traceback:
  PDFParser_LlamaIndex: Unable to extract text from scanned PDF

Analysis:
  This PDF appears to be a scanned image without text layer.

Suggested fixes:
1. Use OCR-enabled parser:
   parsers:
     - type: PDFParser_LlamaIndex
       config:
         use_ocr: true

2. Or use dedicated OCR pipeline:
   See /llamafarm:config for OCR-enabled configuration

Common Log Patterns

Successful Ingestion

[INFO] Processing dataset: my_dataset
[INFO] Parsed 10 documents, 450 chunks created
[INFO] Embedding complete: 450 vectors
[INFO] Indexed to main_db

RAG Query

[INFO] Query: "search terms"
[INFO] Strategy: BasicSimilarityStrategy
[INFO] Retrieved 10 chunks in 0.12s
[INFO] Top score: 0.87

Model Interaction

[INFO] Chat request: model=default, rag=true
[INFO] Context: 5 chunks, 2,340 tokens
[INFO] Response: 512 tokens in 2.1s

Common Errors

Log Levels

Filter by level:

/llamafarm:logs --level ERROR    # Errors and above
/llamafarm:logs --level WARNING  # Warnings and above

Related

• status.md - Overall health check
• start.md - Start services
• stop.md - Stop services