# MED-002: Logging and Error Reporting Implementation

**Status**: ✅ COMPLETE  
**Date**: 2026-02-13  
**Priority**: Medium  
**Effort**: Medium (MED-002)

## Summary

Implemented comprehensive structured logging with JSON formatting, log rotation, error tracking with unique IDs, and user-friendly Discord error messages. All 156 tests pass (15 new tests added).

## Changes Made

### 1. New File: `claude_coordinator/logging_config.py` (371 lines)

**Features:**
- `JSONFormatter`: Structured JSON log output with timestamp, level, module, function, line, message, exception, and custom fields
- `ErrorTracker`: Unique error ID generation and tracking
- `setup_logging()`: Application-wide logging configuration with rotation
- `get_log_directory()`: Smart log directory detection (production vs fallback)
- `format_error_for_discord()`: User-friendly error messages without internal details
- `log_and_format_error()`: Combined logging and message formatting

**Error Types Supported:**
- `timeout`: Claude request timeout
- `claude_error`: Claude CLI failure
- `parse_error`: Malformed JSON response
- `config_error`: Configuration issues
- `permission_error`: Permission denied
- `session_error`: Session management failure
- `lock_error`: Lock acquisition failure

### 2. Updated: `claude_coordinator/bot.py`

**Enhancements:**
- Imported and integrated logging_config module
- Enhanced `main()` to use `setup_logging()` with rotation
- Added structured logging throughout message handling
- Integrated error tracking with unique error IDs
- Enhanced error messages with `log_and_format_error()`
- Added performance metrics (request duration, Claude duration, session operations)
- Added context to all log calls (channel_id, session_id, cost, duration, etc.)

**Key Events Logged:**
- Bot lifecycle (startup, ready, shutdown)
- Message processing (received, matched, session state)
- Claude operations (command, subprocess, response)
- Error handling (with error IDs and context)
- Performance metrics (durations, costs, chunk counts)

### 3. New File: `tests/test_logging.py` (316 lines)

**Test Coverage (15 tests):**
- JSON formatter produces valid JSON
- Required fields in all log entries
- Exception info inclusion
- Extra fields support
- Error ID generation and uniqueness
- Error tracking with context
- Log directory creation
- Handler configuration
- Discord error message formatting
- Privacy protection (no internal details exposed)
- Combined logging and formatting

**All tests pass**: ✅ 156 passed (141 existing + 15 new)

### 4. Updated: `tests/test_bot.py`

Fixed `test_handles_claude_failure` to expect new error message format with error ID instead of internal error details.

### 5. New File: `docs/LOGGING.md`

Comprehensive documentation including:
- Architecture overview
- Component descriptions
- Usage examples
- Log analysis with jq
- Environment variables
- Privacy and security guidelines
- Monitoring recommendations
- Troubleshooting guide

### 6. New File: `test_logging_manual.py`

Manual test script for interactive validation of:
- Basic logging setup
- JSON format validation
- Error tracking
- Discord message formatting
- Log rotation

## Log File Configuration

### Locations
1. **Production**: `/var/log/claude-coordinator/bot.log` (if writable)
2. **Fallback**: `~/.claude-coordinator/logs/bot.log`

### Rotation Settings
- **Max file size**: 10MB
- **Backup count**: 5 rotated files
- **Total retention**: ~50MB (5 × 10MB)
- **Encoding**: UTF-8

### Log Format

**File (JSON)**:
```json
{
    "timestamp": "2026-02-13T19:06:35.545176+00:00",
    "level": "INFO",
    "module": "bot",
    "function": "_handle_claude_request",
    "line": 273,
    "message": "Processing message in channel 12345 for project test-project",
    "channel_id": "12345",
    "project": "test-project",
    "message_length": 150
}
```

**Console (Plain Text)**:
```
2026-02-13 19:06:35,545 - claude_coordinator.bot - INFO - Processing message in channel 12345 for project test-project
```

## Discord Error Messages

Before (exposed internal details):
```
❌ Error running Claude:
```
Command failed: invalid syntax
```
```

After (user-friendly with error ID):
```
❌ **Claude Error**
Something went wrong processing your request.
Please try again or rephrase your message.

_Error ID: `abc12345`_
_Please reference this ID if requesting support._
```

## Testing Results

```bash
$ pytest tests/test_logging.py -v
============================= test session starts ==============================
collected 15 items

tests/test_logging.py::TestJSONFormatter::test_json_formatter_produces_valid_json PASSED
tests/test_logging.py::TestJSONFormatter::test_json_formatter_includes_required_fields PASSED
tests/test_logging.py::TestJSONFormatter::test_json_formatter_includes_exception_info PASSED
tests/test_logging.py::TestJSONFormatter::test_json_formatter_includes_extra_fields PASSED
tests/test_logging.py::TestErrorTracker::test_generate_error_id_returns_string PASSED
tests/test_logging.py::TestErrorTracker::test_generate_error_id_is_unique PASSED
tests/test_logging.py::TestErrorTracker::test_log_error_with_id_includes_error_id PASSED
tests/test_logging.py::TestSetupLogging::test_setup_logging_creates_log_directory PASSED
tests/test_logging.py::TestSetupLogging::test_setup_logging_configures_handlers PASSED
tests/test_logging.py::TestFormatErrorForDiscord::test_format_error_includes_error_id PASSED
tests/test_logging.py::TestFormatErrorForDiscord::test_format_error_handles_known_types PASSED
tests/test_logging.py::TestFormatErrorForDiscord::test_format_error_handles_unknown_types PASSED
tests/test_logging.py::TestFormatErrorForDiscord::test_format_error_does_not_expose_internal_details PASSED
tests/test_logging.py::TestLogAndFormatError::test_log_and_format_error_returns_tuple PASSED
tests/test_logging.py::TestLogAndFormatError::test_log_and_format_error_logs_with_context PASSED

============================== 15 passed in 0.12s
```

```bash
$ pytest -v -m 'not integration'
============================== 156 passed, 1 deselected, 2 warnings in 48.49s ==============================
```

## Environment Variables

- `LOG_LEVEL`: Set logging level (DEBUG, INFO, WARNING, ERROR, CRITICAL)
  - Default: INFO
  - Development: DEBUG
  - Production: INFO

## Usage Examples

### Basic Logging
```python
logger.info(
    "Processing message",
    extra={
        'channel_id': '12345',
        'session_id': 'abc-def',
        'cost_usd': 0.05
    }
)
```

### Error Handling
```python
try:
    result = await operation()
except Exception as e:
    error_id, discord_msg = log_and_format_error(
        logger,
        error_type="claude_error",
        message="Operation failed",
        error=e,
        channel_id=channel_id
    )
    await channel.send(discord_msg)
```

### Performance Tracking
```python
start_time = time.time()
response = await claude_runner.run(...)
duration_ms = int((time.time() - start_time) * 1000)

logger.info(
    "Claude CLI completed",
    extra={
        'duration_ms': duration_ms,
        'cost_usd': response.cost,
        'session_id': response.session_id
    }
)
```

## Log Analysis

```bash
# View real-time logs
tail -f ~/.claude-coordinator/logs/bot.log | jq .

# Find all errors
cat bot.log | jq 'select(.level == "ERROR")'

# Find specific error by ID
cat bot.log* | jq 'select(.error_id == "abc12345")'

# Calculate average Claude CLI duration
cat bot.log | jq 'select(.claude_duration_ms) | .claude_duration_ms' | awk '{sum+=$1; count++} END {print sum/count}'

# Track costs
cat bot.log | jq 'select(.cost_usd) | .cost_usd' | awk '{sum+=$1} END {print "Total: $" sum}'
```

## Privacy Protection

### Logged (Safe)
✅ Channel IDs  
✅ Session IDs (UUIDs)  
✅ Timestamps, durations, costs  
✅ Error messages and stack traces  
✅ Message lengths (not content)

### Never Logged
❌ User message content  
❌ Discord tokens  
❌ Authentication credentials  
❌ Personal identifiable information

### Exposed in Discord
✅ Error type and description  
✅ Error ID for support  
✅ Actionable suggestions

### Never Exposed in Discord
❌ Stack traces  
❌ File paths  
❌ Internal error details  
❌ Session IDs  
❌ Database information

## Monitoring Recommendations

**Real-time Monitoring:**
```bash
# Follow all logs
tail -f ~/.claude-coordinator/logs/bot.log | jq .

# Follow errors only
tail -f ~/.claude-coordinator/logs/bot.log | jq 'select(.level == "ERROR")'
```

**Alerting Criteria:**
- High error rate (>5 errors/minute)
- Repeated timeouts (same channel)
- Session save failures
- Database errors
- Cost spikes (>$1/hour)

## Deployment Notes

1. **Log directory created automatically** on first run
2. **No manual rotation needed** - handled by RotatingFileHandler
3. **Environment variable** `LOG_LEVEL` controls verbosity
4. **Backwards compatible** - no breaking changes to existing functionality

## Validation

✅ All 156 tests pass  
✅ JSON logs parse correctly  
✅ Error IDs are unique  
✅ Discord messages are user-friendly  
✅ Internal details not exposed  
✅ Log rotation works correctly  
✅ Performance metrics captured  
✅ Privacy protection validated

## Benefits

1. **Debugging**: Structured logs with full context make troubleshooting easier
2. **Monitoring**: JSON format enables automated log analysis and alerting
3. **Support**: Error IDs allow users to reference specific issues
4. **Privacy**: Internal details kept out of Discord messages
5. **Performance**: Duration and cost tracking for optimization
6. **Scalability**: Log rotation prevents disk space issues
7. **Observability**: Comprehensive visibility into bot operations

## Next Steps

MED-002 is complete and ready for production deployment.

**Recommended follow-ups:**
- [ ] Monitor logs in production for 1-2 weeks
- [ ] Set up automated alerting for critical errors
- [ ] Create log analysis dashboard
- [ ] Implement external log aggregation (Loki, CloudWatch)
- [ ] Add cost tracking alerts

---

**Files Changed:**
- ✅ `claude_coordinator/logging_config.py` (new, 371 lines)
- ✅ `claude_coordinator/bot.py` (enhanced logging)
- ✅ `tests/test_logging.py` (new, 316 lines, 15 tests)
- ✅ `tests/test_bot.py` (1 test updated)
- ✅ `docs/LOGGING.md` (new, comprehensive docs)
- ✅ `test_logging_manual.py` (new, manual validation)

**Test Results**: 156/156 passed (15 new tests added)  
**Status**: ✅ Ready for deployment