Connection Health & Troubleshooting
Monitor and maintain your integrations
Overview
This guide covers monitoring the health of your connected integrations and troubleshooting common issues.
Integration Health Dashboard
The Integrations dashboard provides a centralized view of all your connections:
- Connection Status: Active/Inactive indicators for each integration
- Last Sync: Timestamp of last successful data synchronization
- API Limits: Usage vs. quota limits for each service
- Error Logs: Recent errors and warnings
- Webhook Logs: Incoming and outgoing webhook events
Access the dashboard via Admin → Integrations.
Common Issues and Solutions
1. Authentication Failures
Symptoms:
- Connection status shows "Inactive"
- Error messages like "Invalid credentials" or "Authentication failed"
- No data sync for extended periods
Causes:
- API keys or OAuth tokens expired or revoked
- Incorrect credentials entered during setup
- Permission changes in the connected service
Solutions:
- Verify credentials in the integration settings
- Re-authenticate the connection
- Check permission scopes in the connected service
- Regenerate API keys if necessary
2. Sync Failures
Symptoms:
- "Last sync" timestamp is outdated
- Data appears incomplete or missing
- Error logs show sync-related errors
Causes:
- Rate limiting or API quota exceeded
- Network connectivity issues
- Changes in the connected service's API
- Data format mismatches
Solutions:
- Check API usage in the service's developer console
- Verify network connectivity to the service
- Review error logs for specific error codes
- Adjust sync frequency or data volume
- Update API versions if needed
3. Webhook Delivery Failures
Symptoms:
- Missing events (calls, SMS, appointments)
- Error logs show webhook delivery failures
- Events appear in the source system but not in af-gemini-connect
Causes:
- Incorrect webhook endpoint URLs
- SSL/TLS certificate issues
- Network firewalls blocking requests
- Payload size limits exceeded
Solutions:
- Verify webhook endpoint URLs are correct
- Check SSL certificates are valid and trusted
- Ensure firewalls allow incoming webhook requests
- Implement retry logic with exponential backoff
- Compress large payloads if needed
4. Rate Limiting
Symptoms:
- Error messages like "Too many requests" or "Rate limit exceeded"
- Sudden cessation of data updates
- 429 HTTP status codes in error logs
Causes:
- Exceeding API rate limits
- Burst traffic patterns
- Shared quota across multiple integrations
Solutions:
- Review rate limit policies in the service's documentation
- Implement request throttling and queuing
- Stagger sync operations to avoid peaks
- Request quota increases if needed
- Use webhooks instead of polling when possible
5. Data Format Changes
Symptoms:
- Sync errors after a service updates its API
- Missing or malformed data fields
- Schema validation failures
Causes:
- Connected service changed their API response format
- Field names or data types changed
- Required fields were added or removed
Solutions:
- Review the service's API changelog
- Update field mappings and transformations
- Implement backward compatibility where possible
- Test changes in a staging environment first
Proactive Monitoring
Set Up Alerts
Configure automated alerts for critical integration events:
- Connection failures — Immediate notification
- Sync failures — Immediate notification
- API quota warnings — When usage exceeds 80% of limit
- Webhook failures — After 3 consecutive delivery failures
Regular Health Checks
Perform routine maintenance:
-
Daily:
- Check the integrations dashboard for any red/yellow indicators
- Review error logs for new entries
- Verify critical syncs completed successfully
-
Weekly:
- Review API usage reports
- Check for service updates or API changes
- Test webhook endpoints with sample payloads
-
Monthly:
- Audit all integration configurations
- Update API keys and refresh tokens
- Review permission scopes and adjust as needed
- Test failover and recovery procedures
Implement Circuit Breakers
Use circuit breaker patterns to prevent cascading failures:
- Closed: Integration is healthy, requests flow normally
- Open: Integration is failing, requests are blocked
- Half-Open: Test phase after recovery, requests limited
This prevents a failing integration from overwhelming your system.
Recovery Procedures
Reconnecting a Failed Integration
- Identify the root cause from error logs
- Fix the underlying issue (update credentials, adjust settings, etc.)
- Test the connection in a staging environment if possible
- Reconnect via the integration settings
- Verify data sync is working correctly
- Monitor closely for 24-48 hours after reconnection
Rolling Back Changes
If a recent change caused integration issues:
- Revert the configuration to the previous working state
- Document the rollback in your change log
- Investigate the issue in a test environment
- Implement a fix and test thoroughly
- Deploy the fix with proper monitoring
Emergency Disconnect
In rare cases, you may need to immediately disconnect an integration:
- Access the integration settings
- Click Disconnect for the problematic integration
- Notify affected users about the service disruption
- Implement manual workarounds if needed
- Begin recovery procedures as soon as possible
Service-Specific Considerations
GoHighLevel (GHL)
- Token expiration: GHL API tokens expire every 60 days
- Webhook endpoints: Ensure your server is publicly accessible
- Rate limits: 100 requests per minute per account
- Field sync: Custom fields may need manual mapping
Google Services
- Quota management: Each Google API has different limits
- Token refresh: OAuth tokens can be long-lived but may need refresh
- Property permissions: Ensure you have admin access to all properties
- API versioning: Be prepared for API version deprecation
Meta Ads
- App review: Required for production use, can take days
- Token expiration: System user tokens expire and need refresh
- Rate limits: Strict limits on ad creation and management
- Business verification: May be required for some permissions
Google Ads
- Manager accounts: Required for managing multiple client accounts
- Token refresh: OAuth tokens need periodic refresh
- Developer token: Required for API access
- Usage limits: Monthly spend limits may affect API access
Getting Help
If you can't resolve an integration issue:
- Check the documentation for the specific service
- Review error logs for specific error codes
- Search the knowledge base for similar issues
- Contact support with detailed information:
- Integration name and ID
- Error messages and codes
- Steps to reproduce
- Screenshots or logs
FAQ
Q: How do I know if an integration is healthy?
A: Check the integrations dashboard for green status indicators and recent sync timestamps.
Q: What should I do if an integration fails?
A: First check the error logs for specific issues, then follow the troubleshooting steps for that integration.
Q: Can I schedule maintenance windows?
A: Yes, you can schedule syncs during off-peak hours to minimize impact.
Q: How do I set up alerts?
A: Configure alert rules in the Admin → System Settings → Alerts section.
Q: What's the most common integration issue?
A: Authentication failures due to expired tokens or changed credentials.
