Bring your towel and join one of the fastest growing data communities. Welcome to our second-generation open source orchestration platform, a completely rethought approach to dataflow automation.

Prefect Community

Hello Team -
After storing/persisting the results of the flows/tasks with the configuration persist_result=True and result_storage=<s3_storage_block>, how can these results be retrieved/managed for later consumption?
Earlier, the endpoint `artifacts/filter` was returning the flowrun metadata and that contains the storage location information. But it is no longer the case now (maybe because results are not artifacts).
Can someone help here, please?

Could anyone answer this question, please?

hi <@U067C7MRF0B> you can interact with the results in your result storage directly (ie using aws cli, boto or otherwise) or you can read the result by storage key like this
```import asyncio

from prefect import task
from prefect.results import ResultStore


@task(
    persist_result=True,
    result_storage_key="my-storage-key",
)
def add_em_up(x: int, y: int) -&gt; int:
    return x + y


async def main():
    add_em_up(42, 42)

    record = await ResultStore().aread("my-storage-key")
    assert record.result == 84, "expected a result of 84"
    print(record.model_dump_json(indent=2))


if __name__ == "__main__":
    asyncio.run(main())```

```» python sandbox/proofs/load_saved_result.py
23:59:33.876 | INFO    | Task run 'add_em_up' - Finished in state Cached(type=COMPLETED)
{
  "metadata": {
    "storage_key": "/Users/nate/.prefect/storage/my-storage-key",
    "expiration": null,
    "serializer": {
      "type": "pickle"
    },
    "prefect_version": "3.1.10.dev2",
    "storage_block_id": null
  },
  "result": 84
}```

I wonder why results are separated from flowrun/taskrun.
Based on your answer, prefect does not preserve any relationship between flowrun/taskrun and its result. What can be a good reason for that?

Thanks <@U02GMEZU18B> for answering my first question. I’ve a followup on that -
In my case, there is a HTTP client seeking the result of a flowrun (after the completion of a flowrun). Is there a way to fetch result through a REST endpoint using the `flowrun id`?

results are not stored on the server, they're stored client side (i.e. in your storage), so there is not a REST endpoint for fetching results

there are many reasons for this, the main one is that results are often sensitive

&gt; there are many reasons for this, the main one is that results are often sensitive
Yes, I understand this <@U02GMEZU18B>
In Prefect v2, `artifacts/filter` endpoint was returning (S3) location of the generated result of a flowrun/taskrun. It seems in Prefect V3 this behaviour has changed and the response from the endpoint does not include result storage location information.
Is there a way to retrieve result storage information (eg: S3 bucket path) of a flowrun/taskrun ?

can you <https://github.com/PrefectHQ/prefect/discussions/new/choose|open a discussion> detailing the change in behavior you've noticed?

<@U02GMEZU18B> - <https://github.com/PrefectHQ/prefect/discussions/16521|here> is the discussion thread on the issue. Please let me know if you need more details