Sometimes, you may want to replicate data to a different location than what you initially connected to Stitch.

For example: You now want to replicate data from your integrations to a different database in your Amazon Redshift cluster, or you simply decide that Amazon Redshift isn’t the destination for you.


The only prerequisite to changing destinations is that the new destination is ready to be connected to Stitch. This will minimize any downtime you may experience.

If you need a refresher on how to spin up a destination for Stitch, check out the destination setup guides.


Here’s what you need to know to ensure a smooth switch:

  • Some destinations may structure data differently than your current destination. For example: If you’re switching from Amazon Redshift to Google BigQuery, there will be some differences in how your data is stored. Detailed info about how Stitch loads data can be found in the Data loading guide for each destination.

  • Your integrations will be paused. All database and SaaS integrations will be automatically paused. After the switch is complete, you’ll need to manually unpause the integrations you’d like to resume.

  • Some webhook data may be lost during this process. Due to webhook’s continuous, real-time nature, some data may be lost.

  • We won’t delete or transfer any data from your current destination. To get historical data into your new destination, you’ll need to queue a full re-replication of all your integrations.

    Re-replicating historical data will count towards your row usage and may take some time, depending on the volume of data and API limitations imposed by the provider.

  • Historical data from webhook-based integrations must be either manually backfilled or replayed. Some webhook providers - such as Segment - allow customers on certain plans to ‘replay’ their historical data. This feature varies from provider to provider and may not always be available.

    If you don’t have the ability to replay historical webhook data, then it must be manually backfilled after the switch is complete.

Switch destinations

To switch to a new destination, you’ll need to:

Step 1: Select a historical data setting

  1. From the Stitch Dashboard, click the Destination tab.
  2. At the bottom of the page, click the Remove Destination button.
  3. In the Connected Source Integrations section, select how you want data to be replicated to the new destination:

    • Replicate new data only: Stitch will pick up where it left off and only replicate new data to your new destination.

    • Replicate historical data: Stitch will clear all Replication Key values, queue a full re-replication of your integrations’ data, and replicate all historical data to your new destination. For SaaS integrations, Stitch will replicate data beginning with the Start Date currently listed in the integration’s settings.

      Note: This isn’t applicable to webhook integrations. Historical data for webhooks must be replayed or manually backfilled after the switch is complete.

Step 2: Delete the current destination

  1. In the Confirm Removal field, enter the display name of the destination.
  2. Click Remove Destination to confirm. You’ll be prompted to confirm the removal of the current destination’s settings.
  3. To continue with the switch, click Remove to delete the current destination settings.

Note: This will not delete data in the destination itself - it only clears this destination’s settings from Stitch.

Step 3: Configure the new destination

  1. On the next page, click the icon of the destination type you want to switch to.
  2. Follow the instructions for that destination type to complete the setup. If you need some help, refer to the destination’s setup guide:

Step 4: Unpause your integrations

After you’ve successfully connected the new destination, un-pause your integrations. Your data will begin replicating according to the historical data option you previously selected.


If you encounter connection issues, check out the Destination Connection Errors guide for common problems and solutions.

Questions? Feedback?

Did this article help? If you have questions or feedback, feel free to submit a pull request with your suggestions, open an issue on GitHub, or reach out to us.