Describe the bug

When using the Iceberg input connector to read a table that contains a timestamptz column, Feldera fails to deserialize the Parquet data with the following error:

Parse error: error deserializing table records from Parquet data:
Error: Unsupported timezone: +00:00 is not supported
Invalid bytes: []

In Apache Iceberg, timestamptz is one of the two standard timestamp types (alongside timestamp). It is extremely common in real-world Iceberg tables — for example, tables synced from
relational databases via JDBC connectors almost always include timestamptz columns for audit fields like created_at or updated_at.

In my test environment,
the Parquet files produced by Iceberg store the timezone as +00:00 (UTC offset notation). Feldera's Parquet reader does not recognize this format, causing the entire table read to fail —
even if the timestamptz column is not included in the Feldera table definition.

To Reproduce

Expected behavior

Screenshots

Context (please complete the following information):

• Feldera Version:
• Environment:
• Browser:
• Pipeline Configuration/SQL code:

Additional context
Add any other context about the problem here.