Troubleshooting

Common issues and solutions for running Flicker.

Common Issues

Logs not appearing in destination

1. Check Process: Ensure Flicker is running (ps aux | grep flicker or systemctl status flicker). Look for startup messages indicating successful config loading.
2. Check Paths: Ensure configured file paths exist and are readable by the user running Flicker.
3. Check Network: Can Flicker reach the destination endpoint? (Try curl -v <endpoint>).
4. Check Destination: Is the destination server receiving requests? Check its access logs.
5. Check Retry Queue: Look for [Retry] messages in Flicker's stdout/stderr. This indicates failed batches are being queued.

High memory usage

• Reduce Buffer Size: Lower buffer_size in your config.
• Reduce Files: If tailing hundreds of files, consider splitting across multiple Flicker instances (though it scales well).
• Check Line Length: Buffers are line-based. Extremely long lines (megabytes) will consume more memory.
• Check Retries: Reduce retry.max_queue_size if destination is down and queues are filling up.

Batches being dropped

• Network Issues: Check connectivity.
• Timeouts: Increase retry.max_retries or retry.max_delay_ms if outages are longer than the default backoff allows.
• Error Messages: Check Flicker logs for specific HTTP or TCP errors.

Missed log entries after restart

• Default Behavior: Without registry tracking, Flicker starts at the end of files to avoid re-shipping old logs.
• Solution: Enable Registry Tracking to persist positions.

File rotation not detected

• Platform: Inode tracking works on Unix/Linux. On Windows, rotation detection is limited.
• Permissions: Ensure Flicker has permission to read the directory to detect new files/inodes.