I have scheduled job to refresh a group of assets. When a job fails, it does not retry. I am deployed on GKE using the K8sRunLauncher. Below is where I am configuring the retries. I am having trouble finding documentation on

define_asset_job

and there is a good chance I am including tags in there that don’t do what I think they should.

assets_schedule = build_schedule_from_partitioned_job(
    define_asset_job(
        "xxxxxx_job",
        selection=AssetSelection.groups("xxxxxx"),
        tags={
            "dagster-k8s/config": {
                "container_config": {
                    "resources": {
                        "requests": {"cpu": "500m", "memory": "1Gi"},
                    }
                },
            },
            "dagster/max_retries": 2,
            "dagster/retry_strategy": "ALL_STEPS"
        },
        partitions_def=daily_partitions,
    ),
    hour_of_day=3,
)

Hey marcos, do you have run retries enabled in the helm chart? https://github.com/dagster-io/dagster/blob/master/helm/dagster/values.yaml#L1046-L1047

@daniel I most certainly did not. Thank you!