API Limit Exceeded in Fivetran – Data Sync Delay (Salesforce)

Overview

You may encounter a data sync delay in Fivetran due to an API limit exceeded error from the source system. This document explains why this occurs, what to verify, and best practices to prevent similar issues in the future.

What Happened?

Fivetran observed that the source system exceeded its allowed API request limits. To prevent breaching hard API limits, Fivetran automatically reschedules the sync until API usage returns to safe levels.

Things to Verify

1. Review API Usage in Fivetran

• Check the Fivetran destination / connector usage.

• Identify which object or table contributed most to the API consumption.

2. Salesforce-Specific Behavior (If Salesforce Is the Source)

Salesforce enforces daily API request limits per organization.

• When API usage reaches 90% of the allowed limit, Fivetran automatically pauses or reschedules syncs.

• This prevents the Salesforce org from fully exhausting its API quota.

• Syncs will resume automatically once:

  • The Salesforce API limit resets, and

  • total_requests_made drops below 90% of total_allowed_requests.

In this case, the Salesforce organization’s API usage was already above the daily limit, which triggered the sync delay.

Information Required from the Customer

To help confirm the root cause, please share the following details:

1. Salesforce Admin Confirmation

   • Were there any recent bulk updates, data imports, or object modifications that could have consumed a large number of API calls?

2. Source System Activity

   • Are there other applications, scripts, or integrations actively updating records (for example, Contact or Case objects) in your Salesforce org?

3. API Usage Evidence (If Available)

   • Screenshots or metrics from Salesforce API usage monitoring showing API consumption trends during the impacted time window.

Is Pausing the Fivetran Connector the Right Approach?

Yes. Pausing the Fivetran connector during bulk Salesforce operations is a valid and recommended best practice.

Why This Works

• Salesforce API limits are shared across the entire org (all apps, integrations, and tools).

• Bulk operations (imports, mass updates, workflow executions) significantly increase API usage.

• Temporarily pausing Fivetran prevents it from competing for API quota.

• This avoids crossing the 90% API threshold, which would otherwise trigger sync rescheduling.

Recommended Best Practices for Future Bulk Operations

1. Before Starting Bulk Updates

• Pause the Fivetran connector (via dashboard or API).

• Perform bulk data loads, imports, or mass updates.

• Ensure all operations complete successfully.

2. After Bulk Updates Complete

• Monitor Salesforce API usage:

  • Salesforce Setup → System Overview → API Usage

• Once API usage returns to normal levels:

  • Resume the Fivetran connector.

• Fivetran will automatically catch up on all changes using the SystemModstamp field.

3. Additional Optimization Options

Adjust Sync Frequency

• Reduce sync frequency during peak business hours.

• Schedule syncs outside known maintenance or bulk processing windows.

Optimize Table and Field Selection

• Review the connector’s Schema tab.

• Disable unnecessary tables or fields not required for analytics.

• This reduces baseline API consumption.

Consider Increasing Salesforce API Limits

• Work with your Salesforce Account Executive to increase daily API limits.

• This provides additional capacity for both operational workloads and Fivetran syncs.

Data Integrity Assurance

No data loss occurs when a connector is paused.

• Fivetran tracks changes using timestamp fields such as SystemModstamp.

• When the connector resumes, all changes made during the pause period are fully synchronized.

Summary

• The sync delay was caused by Salesforce API limits being exceeded.

• Fivetran automatically protects your org by rescheduling syncs.

• Pausing the connector during bulk operations is the recommended prevention strategy.

• Once API usage resets, syncs resume automatically with no data loss.