@kapa.ai I have upgraded to Airbyte 2.0.1. (Deployed on EC2 using

abctl

). I am synchronizing CSV files from S3, but it failed to detect a schema change. I have confirmed directly in the source data that a column has been added. Manually refreshing the schema from the UI also did not detect the change. Has a similar issue been reported? Also, is there a way to work around this?

Based on the knowledge sources provided, there is no explicit mention of a known issue in Airbyte 2.0.1 (or the S3 CSV source connector) where schema changes—such as the addition of a new column—are not detected, even after manually refreshing the schema from the UI. However, here are some relevant points and possible workarounds: 1. Schema Change Detection in Airbyte Airbyte is designed to detect schema changes either automatically (before each sync) or manually (via the UI). For self-managed deployments, schema checks are performed at most every 24 hours, but you can manually refresh the schema at any time from the UI by clicking "Refresh source schema" in the Schema tab of your connection. If there are changes, you should be able to review them in the dialog that appears after the refresh. If the new column is not detected, it suggests either a connector limitation or a bug in schema discovery for your specific setup [Schema change management](https://docs.airbyte.com/platform/using-airbyte/schema-change-management#manually-refresh-the-source-schema). 2. S3 CSV Connector Schema Discovery The S3 source connector (for CSV) infers the schema based on the files it finds. There was a change in version 2.0.0 of the S3 connector to infer schema based on one file instead of all files, which could potentially cause issues if the file used for inference does not contain the new column. This could explain why a new column is not detected if it is not present in the sampled file [S3 Changelog](https://docs.airbyte.com/integrations/sources/s3#changelog). 3. Workarounds - Ensure the new column is present in the first file (or the file used for schema inference) in your S3 bucket, as the connector may only look at one file for schema discovery. - Try re-uploading or renaming files so that the file with the new column is picked up first. - Check for connector updates: If possible, upgrade the S3 connector to a newer version, as there may have been bug fixes or improvements since 2.0.1. - Reset the connection: Sometimes, resetting the connection or clearing the stream can force a full schema re-evaluation. 4. Reporting and Tracking While the knowledge sources do not mention a specific bug report for this issue, similar schema evolution and detection issues have been discussed in the Airbyte community and tracked on GitHub. You may want to check or report your issue on the [Airbyte GitHub issues page](https://github.com/airbytehq/airbyte/issues) or the [community forum](https://discuss.airbyte.io/t/90). Summary: Your experience matches a possible limitation in how the S3 CSV connector infers schema (sampling only one file). Ensuring the new column is present in the sampled file, or resetting the connection, may help. There is no explicit documentation of this as a known bug in 2.0.1, but similar schema evolution issues have been discussed in the community. If these steps do not resolve your issue, consider raising it on the Airbyte GitHub or forum for further assistance.