Troubleshoot the Amplitude connector
This feature is in Beta. Workspace admins can control access to this feature from the Previews page. See Manage Databricks previews.
This page describes common errors you might encounter with the managed Amplitude connector and how to resolve them.
For general troubleshooting guidance that applies to all managed ingestion pipelines, see Troubleshoot managed ingestion pipelines.
The export window is too large
Error message:
The Amplitude export window is too large (exceeded the 4GB limit or timed out). If you set request_time_range, lower it toward the PT2H minimum; if a single window still exceeds 4GB, export that data via the Amazon S3 destination instead.
Cause: A single events export window returned more than 4 GB of data or timed out. Amplitude returns an HTTP 400 when the export size exceeds 4 GB.
Solution:
Lower the request_time_range option for the events table toward the PT2H minimum to shrink each export window. See Connector options. If a single window still exceeds 4 GB, export that data through the Amplitude Amazon S3 destination instead.
The pipeline fails to authenticate
Cause: The API Key or Secret Key stored in the connection is not valid, or the key does not have permission to call the requested endpoint.
Solution:
Confirm that the API Key and Secret Key in the connection match an active project key, and that a Manager or Admin generated them. See Configure authentication to Amplitude. Amplitude does not store secret keys, so if you rotate a key, generate a new secret key and update the connection. See Manage your API keys and secret keys.
Cohorts ingestion hits the request limit
Cause: On Growth and Enterprise plans, the Amplitude Behavioral Cohorts Download API is limited to 500 requests per month. Frequent refreshes of the cohorts table can exhaust this limit.
Solution:
Reduce how often the pipeline refreshes the cohorts table. See Behavioral Cohorts API.