Short answer: - Use

get_run_logger()

inside your

@flow

and

@task

code. It attaches run context (flow/task IDs, names) and automatically ships logs to Prefect Cloud so they show up on the run page. - Use

get_logger()

anywhere outside a run context (module init, setup scripts, infra code). Those logs won’t appear in a specific flow/task run in Cloud unless you add your own handler. Details: -

get_run_logger()

requires an active flow/task run; calling it elsewhere raises an error. It returns a Prefect-aware logger adapter wired to the API handler so logs are sent to Cloud with the right metadata. -

get_logger()

returns a standard Python logger (e.g.,

prefect.*

). By default it logs to console; it’s not tied to a specific run. Examples:

from prefect import flow, task, get_run_logger
from prefect.logging import get_logger

# Outside any run context
bootstrap_logger = get_logger("bootstrap")
<http://bootstrap_logger.info|bootstrap_logger.info>("Starting up...")  # Console; not in a run's logs

@flow
def my_flow():
    logger = get_run_logger()
    <http://logger.info|logger.info>("Flow started")  # Appears in this flow run’s logs in Cloud
    do_work.submit()

@task
def do_work():
    logger = get_run_logger()
    <http://logger.info|logger.info>("Task message")  # Appears in this task run’s logs in Cloud

Notes for your setup (Prefect 2 on Cloud, running on GCP): -

get_run_logger()

logs will be batched and sent to Prefect Cloud by your worker/infra automatically. - If you want third-party/library logs to appear in the run UI, emit them via

get_run_logger()

within the run (or configure logging to propagate appropriately—see guide below). Docs: - get_run_logger API - get_logger API - Logging guide