Here's an example flow:

from prefect import flow, task, get_run_logger

@task
def execute_sql(sql: str):
    logger = get_run_logger()
    <http://logger.info|logger.info>(f"Executing: {sql}")
    <http://logger.info|logger.info>(sql)


@flow
def bracket_flow():
    execute_sql("DROP TABLE [dbo].[SomeTable];")
    execute_sql("EXEC [dbo].[usp_SomeProcedure];")


if __name__ == "__main__":
    bracket_flow()

The output I get is:

18:30:31.900 | INFO    | prefect.engine - Created flow run 'impetuous-sheep' for flow 'bracket-flow'
18:30:32.506 | INFO    | Flow run 'impetuous-sheep' - Created task run 'execute_sql-df2a3f5c-0' for task 'execute_sql'
18:30:32.507 | INFO    | Flow run 'impetuous-sheep' - Executing 'execute_sql-df2a3f5c-0' immediately...
18:30:32.714 | INFO    | Task run 'execute_sql-df2a3f5c-0' - Executing: DROP TABLE .[SomeTable];
18:30:32.715 | INFO    | Task run 'execute_sql-df2a3f5c-0' - DROP TABLE .[SomeTable];
18:30:32.792 | INFO    | Task run 'execute_sql-df2a3f5c-0' - Finished in state Completed()
18:30:32.885 | INFO    | Flow run 'impetuous-sheep' - Created task run 'execute_sql-df2a3f5c-1' for task 'execute_sql'
18:30:32.886 | INFO    | Flow run 'impetuous-sheep' - Executing 'execute_sql-df2a3f5c-1' immediately...
18:30:33.090 | INFO    | Task run 'execute_sql-df2a3f5c-1' - Executing: EXEC .;
18:30:33.091 | INFO    | Task run 'execute_sql-df2a3f5c-1' - EXEC .;
18:30:33.166 | INFO    | Task run 'execute_sql-df2a3f5c-1' - Finished in state Completed()
18:30:33.247 | INFO    | Flow run 'impetuous-sheep' - Finished in state Completed('All states completed.')

The text inside the brackets is being removed in the logs. I thought at first it might be some weird interaction with f-strings, but its happening when I just directly log the variable too.

let me see if I can disable that behavior by default

yeah I’ll make a PR to disable markup by default https://rich.readthedocs.io/en/stable/markup.html#rendering-markup