You are here:
Minimize Impact during Integration Downtime
Learn how to minimize business impact during unexpected integration downtime.
Third-party integrations are common, as they can complement your org. While Salesforce can control and communicate when Salesforce events can cause limited availability, Salesforce can't predict or control third-party downtime. Consider ways to minimize the business impact your third-party integrations have if they were unavailable.
Ensure That Integrations Guarantee Data Integrity and Delivery
Even in the unlikely event that integration middleware or an external system fails during a data update, ensure that all third-party integrations guarantee data integrity and data delivery. Without this assurance, irrecoverable data errors can occur. If discovered too late, the errors can lead to irreversible brand and reputational damage.
- Ensure that every integration call and transaction follows the ACID principle—atomicity, consistency, isolation, durability— even if a service disruption interrupts ongoing transactions. Learn more about the ACID principle in the Understand the Basics of Databases Trailhead.
- For Apex callouts, check the callout’s response, and handle any exceptions.
- For custom Apex web services or Apex REST API, provide appropriate HTTP response codes in your return response to let other systems interpret the transaction outcome.
Monitor Integrations and Diagnose Issues
Establish monitoring that alerts your IT operations team upon discovery of potential integration issues, errors, or failures. To reduce the risk of prolonged deterioration in the user experience or out-of-sync data changes between systems, give your team the ability to diagnose and respond to these discoveries.
- Record any exceptions in your Apex code, and alert appropriate IT stakeholders.
- Debug code in the Developer Console to help uncover potential Apex code issues.
- Use Salesforce monitoring and alerting features, such as Event Monitoring and Proactive Alerts, to review abnormal error behavior.
- Set up monitoring in your integration middleware to discover and diagnose integration failures before they impact your org.
Bulk-Proof the Integration Recovery Process
Create a bulk-proof recovery process, especially for high-volume environments, that efficiently processes the backlog of transactions that accumulated during integration downtime and incoming transactions following integration recovery. A bulk-proof recovery process helps reduce the risk of transaction delays that impact the user experience even after the integration is online.
- Typically, integrations into Salesforce APIs can identify the volume in the backlog at any time. Ideally, integrations can also appropriately switch to a small-volume synchronous processing approach, such as Salesforce REST API, or a large-volume asynchronous processing approach, such asSalesforce Bulk API, depending on the backlog volume.
- In the unlikely event that the backlog volume grows too large, ensure that other systems in the integration can process records automatically in bulk or manually.
- If possible, design solutions to process records in bulk, not just one record at a time, so the system can effectively handle records in large volumes in batches.
