Assuming you’re on Prefect 3.x: every Task Run has a

cache_key

computed from its cache policy. You can retrieve it a few ways: - Inside the task while it’s running

from prefect import task, get_run_context, get_run_logger

@task
def my_task(x):
    ctx = get_run_context()          # TaskRunContext
    logger = get_run_logger()
    <http://logger.info|logger.info>(f"cache_key={ctx.task_run.cache_key}")
    return x * 2

- From the API using the Python client (given a task run ID)

import asyncio
from prefect import get_client

async def get_cache_key(task_run_id):
    async with get_client() as client:
        tr = await client.read_task_run(task_run_id)
        return tr.cache_key

# asyncio.run(get_cache_key(UUID("...")))

- From a PrefectFuture (to get the task run ID first)

from prefect import flow

@flow
def parent():
    fut = my_task.submit(1)
    print(f"task_run_id={fut.task_run_id}")

Then pass that

task_run_id

to the client snippet above. - Via REST API directly Make a GET request to

/task_runs/{id}

and read the

cache_key

field: Task run by id Notes - The Prefect UI does not currently surface

cache_key

directly; use the API or log it from within the task. - For how keys are computed and how to customize them, see Caching concepts. If you’re on Prefect 2.x, the pattern is similar (the

cache_key

is still on the Task Run), but let me know and I can tailor the examples. Do you want to fetch the key for an in-flight run, a past run, or compute what it would be ahead of time?