Update Error Handling to Reflect Accurate Error Message

Description

When data inserted via sender.dataframe() exceeds the configured max_buffer_size, calling sender.flush() raises an exception with the message "All values are null." — which is misleading and does not reflect the actual root cause (buffer overflow).

The root cause is only visible via err.__cause__, which shows a different, more accurate error. This makes debugging significantly harder for users.

Steps to reproduce

import questdb.ingress as qi

with qi.Sender(..., max_buffer_size=<small_value>) as sender:
    try:
        sender.dataframe(df, table_name="my_table")
        sender.flush()
    except Exception as err:
        print(err)          # Misleading: "All values are null."
        print(err.__cause__)  # Accurate: buffer size exceeded

Expected behavior

The top-level exception message should accurately describe the failure — e.g., "Buffer size exceeded: data exceeds the configured max_buffer_size limit." The root cause should not need to be inspected separately to understand what went wrong.

Actual behavior

err prints: All values are null.
err.__cause__ prints: the real buffer overflow error

Environment

• Language: Python
• Client: QuestDB Python ingress client (questdb.ingress)
• Method: sender.dataframe() + sender.flush()

Suggested fix

The exception raised by flush() when the buffer limit is exceeded should surface the real cause directly in its message, or chain exceptions in a way that makes the root cause immediately visible without inspecting __cause__ manually.

Start at the Python client's sender.dataframe() and sender.flush() flow, reproducing the issue with a small max_buffer_size and the example exception handling. Trace how the buffer-overflow error becomes the top-level "All values are null." message. Done means flush() directly exposes an accurate buffer-limit error, with coverage for the dataframe and flush scenario.