Troubleshooting
This comprehensive troubleshooting guide helps you resolve common issues with the FEL AI Product Advisor plugin.
Before diving into specific issues, try these general troubleshooting steps:
- Clear cache:
bin/console cache:clear - Rebuild assets:
bin/build-storefront.sh - Check browser console (F12) for JavaScript errors
- Verify OpenAI API status: https://status.openai.com
Chatbot Not Appearing
Issue: Chatbot element not visible on the page
Possible Causes & Solutions:
1. Plugin Not Activated
Check:
# List installed plugins
bin/console plugin:list | grep FelAIProductAdvisor
Solution:
- Navigate to Settings → System → Plugins
- Find "FEL AI Product Advisor"
- Ensure toggle is activated (green)
- Clear cache:
bin/console cache:clear
2. API Key Missing or Invalid
Check:
- Navigate to Settings → Plugins → FEL OAI Assistants Manager
- Click "..." → Configuration
- Verify API key is entered
- Click "API Test" button to validate
Solution:
- Obtain valid API key from OpenAI Platform
- Re-enter the key and save
- Test the connection
3. Assistant Not Selected
Check:
- Edit the CMS page with the chatbot element
- Click on the chatbot element
- Verify "Assistent auswählen" field has a value
Solution:
- Select a configured OpenAI Assistant from the dropdown
- If no assistants available, create one in the base plugin first
4. Assets Not Compiled
Check:
- Browser console shows 404 errors for JS/CSS files
- Chatbot appears broken or unstyled
Solution:
# Rebuild storefront assets
bin/build-storefront.sh
# Clear cache
bin/console cache:clear
5. Browser Cache Issues
Solution:
- Hard refresh: Ctrl + F5 (Windows) or Cmd + Shift + R (Mac)
- Clear browser cache completely
- Try incognito/private browsing mode
Chatbot Not Responding
Issue: Chatbot displays but doesn't answer messages
Possible Causes & Solutions:
1. OpenAI API Connection Issues
Check:
- Browser console (F12) → Network tab
- Look for failed API requests to OpenAI
- Check response errors
Solution:
- Verify API key is valid and has credits
- Check OpenAI API status: https://status.openai.com
- Ensure server can reach
https://api.openai.com - Verify firewall/proxy allows outbound HTTPS
2. Insufficient OpenAI Credits
Check:
- Log into OpenAI Platform
- Navigate to Billing → Usage
- Verify account has available credits
Solution:
- Add credits to OpenAI account
- Set up payment method
- Check spending limits
3. Assistant Configuration Error
Check:
- Navigate to base plugin: FEL OAI Assistants Manager
- Verify Assistant is properly configured
- Check Assistant has necessary functions enabled
Solution:
- Reconfigure Assistant in base plugin
- Ensure required functions are activated
- Test Assistant independently
4. JavaScript Errors
Check:
- Open browser console (F12)
- Look for red error messages
- Check for conflicts with other plugins
Solution:
# Rebuild administration
bin/build-administration.sh
# Rebuild storefront
bin/build-storefront.sh
# Clear all caches
rm -rf var/cache/*
bin/console cache:clear
Performance Issues
Issue: Chatbot is very slow to respond
Optimization Strategies:
1. Reduce Data Complexity
Action:
- Limit Categories: Exclude unnecessary categories in plugin settings
- Filter Properties: Select only relevant property groups
- Reduce Description Length: Lower max description length (default: 500)
- Limit Search Results: Reduce search result limit (default: 8)
Configuration Location: Settings → Plugins → FEL AI Product Advisor → Configuration
2. Optimize API Settings
Adjust in base plugin configuration:
| Setting | Recommended Value | Impact |
|---|---|---|
| Wartezeit (Sekunden) | 1-2 | Lower = faster, but may timeout |
| Max. Versuche | 30-40 | Lower = faster timeout |
Don't set wait time below 1 second - may cause API errors!
3. Use Faster AI Model
Model Performance Comparison:
| Model | Speed | Cost | Quality |
|---|---|---|---|
| GPT-4o-mini | ⚡⚡⚡ Very Fast | € Very Low | ⭐⭐⭐⭐ Excellent |
| GPT-5-mini | ⚡⚡⚡ Very Fast | € Very Low | ⭐⭐⭐⭐ Excellent |
| GPT-4o | ⚡⚡ Fast | €€ Medium | ⭐⭐⭐⭐⭐ Outstanding |
| Reasoning Models | ⚡ Moderate | €€€ Higher | ⭐⭐⭐⭐⭐ Outstanding |
Recommendation: Start with GPT-4o-mini or GPT-5-mini for optimal speed and cost efficiency
4. Optimize Product Data
Best Practices:
- Keep product descriptions concise
- Reduce number of product variants
- Optimize category structure
- Minimize custom fields
Configuration Issues
Issue: Checkboxes not visible in plugin configuration (Shopware 6.7)
Cause:
Plugin uses modern Shopware 6.7 type="bool" fields
Solution:
# Rebuild administration
bin/build-administration.sh
# Clear cache
bin/console cache:clear
Then refresh browser (Ctrl + F5)
Issue: Theme colors not applying
Check:
- ✅ Standard (Anpassbar) theme selected?
- Other themes have fixed colors
- ✅ Hex color format correct?
- Must be format:
#3b82f6(include the#)
- Must be format:
- ✅ Storefront compiled after change?
Solution:
# Rebuild storefront
bin/build-storefront.sh
# Clear cache
bin/console cache:clear
Hard refresh browser: Ctrl + F5
Issue: CMS element not available in editor
Check:
- Plugin activated?
- Assets compiled?
- Browser cache cleared?
Solution:
# Refresh plugin list
bin/console plugin:refresh
# Verify activation
bin/console plugin:list | grep FelAIProductAdvisor
# Rebuild storefront
bin/build-storefront.sh
# Clear cache
bin/console cache:clear
Error: "API Key Invalid"
Cause:
- Invalid or expired OpenAI API key
- API key not properly saved
Solution:
- Navigate to FEL OAI Assistants Manager plugin settings
- Re-enter your OpenAI API key
- Click "API Test" to verify
- Save configuration
- Clear cache:
bin/console cache:clear
Error: "Rate Limit Exceeded"
Cause:
- Too many API requests in short time
- OpenAI rate limits reached
Solution:
- Wait a few minutes before trying again
- Increase Wartezeit (Sekunden) in configuration
- Upgrade OpenAI plan for higher rate limits
- Reduce concurrent chatbot usage
Error: "Assistant Not Found"
Cause:
- Selected Assistant was deleted in OpenAI
- Assistant ID mismatch
Solution:
- Navigate to base plugin: FEL OAI Assistants Manager
- Verify Assistant still exists
- Recreate Assistant if deleted
- Re-select Assistant in CMS element configuration
Debugging
Enable Shopware Debug Mode
For detailed error information:
In .env.local:
APP_ENV=dev
Then:
bin/console cache:clear
Never leave debug mode enabled on production systems!
Check Shopware Logs
Monitor logs for errors:
# Real-time log monitoring
tail -f var/log/prod-*.log
# Search for specific errors
grep -i "error" var/log/prod-*.log
# Check recent entries
tail -n 100 var/log/prod-*.log
Browser Console Debugging
Enable browser developer tools:
- Press F12 (or Cmd + Option + I on Mac)
- Go to Console tab
- Look for red error messages
- Go to Network tab to check API requests
Common Console Errors:
| Error | Cause | Solution |
|---|---|---|
404 Not Found | Assets not compiled | Rebuild storefront |
Uncaught TypeError | JavaScript conflict | Check other plugins |
Network Error | API unreachable | Check firewall/connectivity |
CORS Error | Cross-origin issue | Verify API configuration |
Common Questions & Solutions
Q: Chat history disappears on page reload
A: Chat history is stored in browser's LocalStorage. This is intentional for privacy.
Options:
- Chat persists during same session
- Closes when browser/tab closes
- No server-side storage by design
Q: Chatbot shows products from wrong category
A: Check category restriction settings:
- Edit CMS element
- Verify "Kategorie" field
- Ensure correct category selected
- Check "Kategorien ausschließen" in global settings
Q: Product images not displaying in chat
A: Possible causes:
Image URL issues
- Check product has media assigned
- Verify media URLs are accessible
- Check file permissions
Thumbnail generation
# Regenerate thumbnails
bin/console media:generate-thumbnails
Q: Multiple chatbots on one page conflict
A: This should work normally, but to avoid issues:
- Use different Assistants for each instance
- Ensure unique Berater Titel for each
- Test thoroughly before deploying
Performance Optimization Checklist
Use this checklist to diagnose performance issues:
- ✅ Using GPT-4o-mini or GPT-5-mini (fastest, most cost-effective models)
- ✅ Excluded unnecessary categories
- ✅ Limited property groups (< 10)
- ✅ Max description length set to 300-500
- ✅ Search result limit: 6-8 products
- ✅ Display limit: 4 products in chat
- ✅ Wartezeit set to 1-2 seconds
- ✅ Max. Versuche set to 30-40
- ✅ Product descriptions are concise
- ✅ Server has adequate resources (2GB+ RAM)
Getting Help
Before Contacting Support
Gather this information:
System Information:
- Shopware version:
bin/console --version - PHP version:
php -v - Plugin version: Check in plugin manager
- Shopware version:
Error Details:
- Exact error message
- When error occurs
- Steps to reproduce
- Browser console screenshot
Log Files:
- Recent entries from
var/log/prod-*.log - JavaScript console errors
- Network tab errors (F12)
- Recent entries from
Contact Support
Email:info@5-elements-design.de
Include in your message:
- System information (above)
- Detailed error description
- Screenshots of errors
- Steps to reproduce
- What you've already tried
Response Time:
- Standard: 24-48 hours
- Urgent: Mark email subject with [URGENT]
Additional Resources
- OpenAI Status: https://status.openai.com
- Shopware Documentation: https://developer.shopware.com
- OpenAI API Docs: https://platform.openai.com/docs
- Community Forum: Contact support for access
FAQ: Frequently Asked Questions
Q: Does the plugin cost money to use? A: The plugin itself may have a license cost, but you also need an OpenAI API account with credits. Costs are per API request. With mini models (GPT-4o-mini, GPT-5-mini), API costs are very affordable. Premium models require monitoring token usage.
Q: How many Assistants can I create? A: Unlimited - create different Assistants for different categories or languages.
Q: Can I have multiple chatbots on one page? A: Yes, you can add multiple CMS elements with different Assistant configurations.
Q: Does it support multiple languages? A: Yes, the chatbot automatically adapts to the customer's language. Create separate Assistants per language for best results.
Q: Are chat conversations saved? A: Locally in the customer's browser (LocalStorage) only. No server-side persistent storage for privacy reasons.
Q: What happens if OpenAI is down? A: The chatbot won't function during outages. Check https://status.openai.com for status updates.
If you couldn't find a solution here, don't hesitate to contact our support team at support@5elements.de. We're here to help!