Release Notes

May 2026

We have added a new table, MERGED_USER.

April 2026

We've implemented a new mechanism to fetch source schema quicker when creating a new connection. Now, instead of deriving the schema from source data, we use fixed schema models. This allows you to see the schema blocking and hashing screen almost immediately after creating a new connection.

As a result of this improvement, you may see empty tables and columns in the Schema tab of existing connections that have not yet synced to your destination.

After optimizing schema fetch for the connectors below, these connections no longer use the pre-initial-sync schema capture flow described in schema_status.

Affected connectors

REST API request changes (breaking)

For the connectors listed above, the schema_status request field is no longer supported in:

Update a Connection

If your automation still sends schema_status for these connectors, requests may fail.

REST API response changes

For the connectors listed above, schema_status is no longer returned for flows that previously exposed it, including:

If your downstream transformations or orchestration rely on this field, update them to handle its absence.

Dashboard flow change

In the Fivetran dashboard, these connectors no longer show the schema prefetch and review step before initial sync. After connection creation, you proceed directly to the initial sync flow.

October 2025

We unified each connector’s pre-built dbt Core–compatible data models into a single model. It standardizes and documents the Fivetran schemas created in your destination and outputs analytics-ready tables. Previously, most connectors used distinct source and transform models.

July 2025

We no longer support the AWS S3 bucket export mechanism.

March 2024

Fivetran now supports extracting event data from Amplitude using Cloud Storage Export. We are gradually rolling out this feature to all existing connections.

For more information, see our Amplitude setup guide.

March 2023

We can now sync your data from Amplitude sources hosted in the EU region.

We now insert null values for the following columns to support their deprecation:

amplitude_event_type
device_brand
device_manufacturer
device_model
is_attribution_event
user_creation_time

However, these columns continue to exist in the destination schema. For more information on the changes, see Amplitude’s community forum.

October 2022

We have released pre-built, dbt Core-compatible data models for Amplitude. Find the models in Fivetran's dbt hub. Learn more in our Fivetran Data Models documentation.

August 2022

We have changed the composite primary key of the EVENT table. The new composite key is a combination of the server_upload_time, event_id, and device_id parameters.

May 2021

We now obfuscate the apiKey and secretKey values in the connector setup form.

December 2020

You can now exclude the EVENT_AMPLITUDE_ATTRIBUTION_ID table from the sync. On your connector details page, go to the Schema tab and deselect the table. By default, we include the EVENT_AMPLITUDE_ATTRIBUTION_ID table in the sync. If you exclude the table and later want to sync the table, perform a complete table re-sync to ensure data integrity.

April 2020

You can now configure your Amplitude connector using the Fivetran REST API. This feature is in BETA and is available only for Standard and Enterprise accounts.

December 2019

We no longer deliver a separate column for each custom field in the EVENTS table. The EVENTS table now syncs custom fields as JSON objects in the following columns:

data
event_properties
groups
group_properties
user_properties

Storing custom fields as JSON objects speeds up syncs and avoids hitting column limits in your destination.

We will discontinue the separate custom field columns after three months, so adjust your queries accordingly. You must also re-sync your EVENTS data during this time so that all your custom field data is in the new format before the separate columns are deprecated.

November 2019

The Amplitude connector now supports priority-first sync. Priority-first syncs fetch your most recent data first so that it's quickly ready for you to use. Amplitude's priority-first syncs fetch the most recent 15 days' worth of data from all tables.

To learn more, see our priority-first sync documentation.

May 2019

We have added a new column, user_id, to the EVENT table.

April 2019

We have deleted the USER table and all its dependent tables:

USER_ALIAS_ID
USER_AMPLITUDE_ID
USER_ALIASING_PROFILE
USER_PROPERTY
USER_MERGE_TIME

All the information in these tables exist in the EVENT table.

March 2019

We have added three new tables:

USER_ALIASING_PROFILE
USER_PROPERTY
USER_MERGE_TIME

We have changed the composite primary key of the EVENT table. The new primary key consists of the event_id, device_id, and client_event_time columns. Previously, the key consisted of the event_id and projectName columns.

We have deleted the DEVICE table.

We have changed the value of initial_sync_start_time to 2014-07-16. This change means that for most Amplitude tables (except EVENT_TYPE), we only sync data that was created after July 16, 2014. We have made this change to protect the integrity of your data.

Amplitude introduced the client_event_time parameter, a part of the primary key for the EVENT table, on July 16, 2014. Because earlier events don’t have the client_event_time parameter, we can’t form the primary key for them. Without the primary key, we cannot deduplicate the data. Therefore, we cannot write that data to the destination if that data was created before Amplitude introduced the client_event_time parameter.

February 2019

We have added a new table, EVENT_AMPLITUDE_ATTRIBUTION_ID.

We have changed the primary key of the USER table from user_id to amplitude_id and project_name.

January 2019

We have added a new table, EVENT_AMPLITUDE_ATTRIBUTION_ID.

Amplitude is a comprehensive product analytics software for web and mobile.

Read our Amplitude connector documentation.