I'm running hybrid setup (Dagster Cloud + K8s). I'm doing this query inside an asset, and the data it returns is wrong (includes runs from other jobs, includes runs outside past 12 hours). The idea is to create an asset with the failed partitions after X number of retries. Any ideas?

context.instance.get_run_partition_data(
        runs_filter=RunsFilter(
            statuses=[DagsterRunStatus.FAILURE, DagsterRunStatus.SUCCESS],
            job_name=TARGET_JOB_NAME,
            updated_after=datetime.now() - timedelta(hours=12),
        )
    )

Hi Juan. Thanks for reporting this. This is definitely a bug - we’ll try to get this patched up ASAP (most likely tomorrow).