Debugging Guide

# Standard output logs
/tmp/agenticai/agenticai.log

# Error logs
/tmp/agenticai/agenticai-error.log

# Follow standard logs
agenticai service logs -f

# Follow error logs only
tail -f /tmp/agenticai/agenticai-error.log

# View last 50 lines
agenticai service logs -n 50

2024-12-18T10:15:23.456789 [info     ] Starting call manager
2024-12-18T10:15:24.123456 [info     ] Call initiated call_id=abc-123 to_number=+1234567890
2024-12-18T10:15:25.789012 [error    ] WebSocket connection failed error=Connection refused

agenticai status

agenticai test-connection

Testing Twilio...
  ✓ Twilio OK
Testing Gateway...
  ✓ Gateway OK
Checking Gemini config...
  ✓ Gemini configured

┏━━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━┓
┃ Service ┃ Status       ┃ Details              ┃
┡━━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━┩
│ Twilio  │ ✓ Connected  │ Account: My Account  │
│ Gateway │ ✓ Connected  │ URL: ws://127.0.0.1  │
│ Gemini  │ ✓ Configured │ Model: gemini-2.5... │
└─────────┴──────────────┴──────────────────────┘

# Check if service is installed
agenticai service status

# View service configuration
cat ~/Library/LaunchAgents/com.agenticai.server.plist

# Check if process is running
ps aux | grep agenticai

# View service PID
launchctl list | grep agenticai

╭─ Agentic AI Service Status ─╮
│ ✓ Service installed          │
│   Plist: ~/Library/...       │
│ ✓ Service running (PID: 1234)│
│                              │
│ Recent logs:                 │
│   Starting call manager      │
│   Call manager started       │
│   Server listening on 8080   │
╰──────────────────────────────╯

# Check server health
curl http://localhost:8080/health

# Expected response
{"status":"healthy","active_calls":0}

# Search for connection errors
grep -i "connection\|failed\|error" /tmp/agenticai/agenticai.log

# Common error patterns:
[error    ] Failed to connect to OpenAI Realtime
[error    ] WebSocket connection failed error=Connection refused
[error    ] Invalid webhook URL

# Filter audio-related logs
grep -i "audio\|bridge\|twilio\|openai" /tmp/agenticai/agenticai.log

# Key indicators:
=== AUDIO BRIDGE STARTED ===
=== TWILIO WEBSOCKET CONNECTED ===
=== OPENAI REALTIME CONNECTED ===

# Search for auth errors
grep -i "auth\|unauthorized\|invalid.*key\|403\|401" /tmp/agenticai/agenticai-error.log

# Examples:
[error    ] Twilio authentication failed
[error    ] OpenAI API key invalid
[error    ] Unauthorized: Invalid API key

# Track a specific call
grep "call_id=abc-123" /tmp/agenticai/agenticai.log

# Expected flow:
[info] Initiating call call_id=abc-123
[info] Call status changed call_id=abc-123 new_status=ringing
[info] Voice webhook called call_sid=CA123
[info] Media stream handler started
[info] Session ended call_id=abc-123 duration=45.2

# Add debug prints in your code
print(f"=== WebSocket message: {data} ===", flush=True)

grep "websocket\|media-stream" /tmp/agenticai/agenticai.log

# Key events:
event=connected    # WebSocket established
event=start        # Stream started
event=media        # Audio data flowing
event=stop         # Stream ended

# Test API key
curl https://api.openai.com/v1/models \
  -H "Authorization: Bearer $OPENAI_API_KEY" | jq '.'

# Check for Realtime API access
curl https://api.openai.com/v1/models/gpt-4o-realtime-preview-2024-12-17 \
  -H "Authorization: Bearer $OPENAI_API_KEY"

grep -i "openai\|realtime" /tmp/agenticai/agenticai.log

# Interactive gateway terminal
agenticai bot

# Commands to test:
openclaw> status
openclaw> ping
openclaw> config

grep -i "gateway\|clawdbot" /tmp/agenticai/agenticai.log

# Expected patterns:
[info] Connecting to OpenClaw Gateway url=ws://127.0.0.1:18789
[info] Gateway connected
[info] Sending message to gateway type=call_started

# Test bot API
curl "https://api.telegram.org/bot$TELEGRAM_BOT_TOKEN/getMe"

# Send test message
curl -X POST "https://api.telegram.org/bot$TELEGRAM_BOT_TOKEN/sendMessage" \
  -d "chat_id=$TELEGRAM_CHAT_ID" \
  -d "text=Test message from Agentic AI"

grep -i "telegram" /tmp/agenticai/agenticai.log

# Set debug environment variable
export PYTHONVERBOSE=1
export PYTHONDEBUG=1

# Run server with verbose output
agenticai server

# Install tcpdump or use Wireshark
sudo tcpdump -i any -n port 8080 -w agenticai-traffic.pcap

# Monitor specific connections
sudo tcpdump -i any -n host api.openai.com

# Find stack traces in error log
grep -A 20 "Traceback" /tmp/agenticai/agenticai-error.log

[error] Error in media stream error=Connection reset exc_info=True
Traceback (most recent call last):
  File "agenticai/core/call_manager.py", line 280
  File "agenticai/twilio/websocket.py", line 45
ConnectionResetError: [Errno 54] Connection reset by peer

# Add to any file
import pdb; pdb.set_trace()

# Or use ipdb for better experience
import ipdb; ipdb.set_trace()

agenticai server  # Breakpoint will pause execution

# Show only errors
grep "\[error" /tmp/agenticai/agenticai.log

# Show warnings and errors
grep -E "\[(error|warn)" /tmp/agenticai/agenticai.log

# Filter by call ID
grep "call_id=abc-123" /tmp/agenticai/agenticai.log

# Show last hour of logs
find /tmp/agenticai -name "*.log" -mmin -60 -exec tail -f {} \;

# Count error types
grep "\[error" /tmp/agenticai/agenticai.log | cut -d']' -f2 | sort | uniq -c

# Check log sizes
du -sh /tmp/agenticai/*.log

# Truncate logs
> /tmp/agenticai/agenticai.log

# Archive old logs
mv /tmp/agenticai/agenticai.log /tmp/agenticai/agenticai.log.$(date +%Y%m%d)

# Check CPU and memory usage
top -pid $(pgrep -f agenticai)

# Or use htop
htop -p $(pgrep -f agenticai)

# Monitor network connections
lsof -i -P | grep agenticai

# Look for slow operations in logs
grep -i "slow\|timeout\|took.*ms" /tmp/agenticai/agenticai.log

# Time specific operations
time agenticai trigger --to +1234567890

# Gather system info
echo "Python: $(python --version)"
echo "OS: $(uname -a)"
echo "Agentic AI: $(agenticai --version 2>/dev/null || echo 'unknown')"
pip list | grep -E "fastapi|twilio|openai"

# Show config without secrets
cat config.yaml | sed 's/api_key:.*/api_key: [REDACTED]/' \
                | sed 's/auth_token:.*/auth_token: [REDACTED]/'

# Extract last 100 lines with context
tail -n 100 /tmp/agenticai/agenticai-error.log > error-report.log

agenticai service start
agenticai trigger --to +1234567890 --webhook-url https://example.ngrok.io
# Error occurs here

# 1. Verify server is listening
curl http://localhost:8080/health

# 2. Check webhook accessibility
curl https://your-url.ngrok.io/health

# 3. Test Twilio credentials
agenticai test-connection

# 4. Monitor logs during call
agenticai service logs -f &
agenticai trigger --to +1234567890

# 1. Check OpenAI configuration
grep "openai_realtime" config.yaml

# 2. Verify API key
echo $OPENAI_API_KEY | cut -c1-10

# 3. Look for audio errors
grep -i "audio\|openai\|realtime" /tmp/agenticai/agenticai-error.log

# 4. Check WebSocket connection
grep "WEBSOCKET CONNECTED" /tmp/agenticai/agenticai.log

# 1. Check crash logs
tail -n 50 /tmp/agenticai/agenticai-error.log

# 2. Test manual start
agenticai server  # Watch for immediate errors

# 3. Check service configuration
cat ~/Library/LaunchAgents/com.agenticai.server.plist

# 4. Verify dependencies
pip check