# Best Practices ## Recommended setup


Whitelist IPs before connecting	Add Coupler.io's IPs (52.21.222.113 and 35.170.113.181) to your firewall or security group before you create a data flow. This prevents "connection refused" errors and saves troubleshooting time.
Use a dedicated database user	Create a MySQL user specifically for Coupler.io with SELECT-only permissions. This limits the blast radius if credentials are ever compromised and makes audit logs cleaner.
Test with a small table first	Before exporting your largest or most critical table, test the connection with a smaller table to confirm everything works smoothly.

## Data refresh and scheduling


Filter by date for large tables	If exporting millions of rows, apply a date filter (e.g., last 30 days) instead of exporting all history. This speeds up exports and reduces timeout risk. You can append historical data in a separate flow if needed.
Start with daily refreshes	Run a daily schedule first to ensure exports complete reliably. Once confident, increase frequency to hourly if your business requires real-time data.
Monitor your first few runs	After scheduling a data flow, check your destination (Google Sheets, BigQuery, etc.) to confirm data is arriving correctly and in the expected format before moving on.

## Performance optimization


Index your filtered columns	If filtering on created_at, status, or user_id, ensure those columns have database indexes. This dramatically speeds up query execution and prevents timeouts.
Export views, not complex queries	If you need complex transformations, create a MySQL view and export that instead of filtering in Coupler.io. Views are pre-computed and run faster.

## Common pitfalls {% columns %} {% column %} **Do** * Whitelist IPs in your firewall or security group before connecting * Create a dedicated MySQL user with SELECT-only permissions * Apply date filters to large tables to avoid timeouts * Test your first export with a manual run before scheduling * Use indexes on columns you filter by (created\_at, status, user\_id) {% endcolumn %} {% column %} **Don't** * Use your root or admin MySQL account — create a separate user * Export your entire database history in one flow if it's millions of rows * Skip firewall configuration if your database is in the cloud or behind a corporate network * Schedule very high frequency refreshes (hourly) for large tables without testing first * Store your MySQL password in plain text notes; use a password manager {% endcolumn %} {% endcolumns %}