apache/beam · error · RuntimeError

Could not find the feature service

Error message

Could not find the feature service %s for the feature store configured in %s.

What it means

Raised in `__enter__` when `store.get_feature_service(self.feature_service_name)` raises, i.e. the configured feature service name does not exist in the Feast registry described by the yaml. The handler converts this into a RuntimeError naming both the service and the yaml path.

Solutions

List available services (`store.list_feature_services()`) locally and use the exact name.
Sync/refresh the registry file the yaml points to so it contains the feature service.
Confirm the yaml's `project` matches the project where the feature service was defined.
Fix typos in `feature_service_name`.

Example fix

// before
feature_service_name='user_daily_features_v1'
// after (name in registry)
feature_service_name='user_daily_features'

Defensive patterns

Strategy: validation

Validate before calling

from feast import FeatureStore
store = FeatureStore(fs_yaml_file='feature_store.yaml')
assert feature_service_name in {fs.name for fs in store.list_feature_services()}

Try / catch

try:
    svc = store.get_feature_service(name)
except Exception as e:
    raise RuntimeError(f'Feature service {name} not found: {e!r}') from e

Prevention

Regenerate the registry file after any feature repo change.
Confirm the feast project in the yaml matches the feature service's project.
Keep feature service names in shared constants between repo and pipeline.

When it happens

Trigger: Constructing the handler with a `feature_service_name` string that is absent from the feature store registry — renamed service, wrong Feast project, stale registry file, or a typo.

Common situations: Feature service renamed in the Feast repo without updating the Beam pipeline config; registry.db rebuilt and the service removed; pointing the yaml at a different project than the one containing the service.

Understand the failure class

Background: "Not found" and "does not exist" errors: why "Task not found", "No such folder", and "Can't find" fire when a lookup comes back empty — this error's family across 14 libraries.

Related errors

AI-assisted analysis of apache/beam@12126d8942 (2026-09-13). Data as JSON: /api/errors/8b7b81ced58905cc. Report an issue: GitHub.

Appendix: source

Thrown at sdks/python/apache_beam/transforms/enrichment_handlers/feast_feature_store.py:160

    _validate_feature_store_yaml_path_exists(self.feature_store_yaml_path)
    _validate_feature_names(self.feature_names, self.feature_service_name)

  def __enter__(self):
    """Connect with the Feast feature store."""
    local_repo_path = download_fs_yaml_file(self.feature_store_yaml_path)
    try:
      self.store = FeatureStore(fs_yaml_file=local_repo_path)
    except Exception:
      raise RuntimeError(
          'Invalid feature store yaml file provided. Make sure '
          'the %s contains the valid configuration for Feast feature store.' %
          self.feature_store_yaml_path)
    if self.feature_service_name:
      try:
        self.features = self.store.get_feature_service(
            self.feature_service_name)
      except Exception:
        raise RuntimeError(
            'Could not find the feature service %s for the feature '
            'store configured in %s.' %
            (self.feature_service_name, self.feature_store_yaml_path))
    else:
      self.features = self.feature_names

  def __call__(self, request: beam.Row, *args, **kwargs):
    """Fetches feature values for an entity-id from the Feast feature store.

    Args:
      request: the input `beam.Row` to enrich.
    """
    if self.entity_row_fn:
      entity_dict = self.entity_row_fn(request)
    else:
      request_dict = request._asdict()
      entity_dict = {self.entity_id: request_dict[self.entity_id]}
    feature_values = self.store.get_online_features(

View on GitHub (pinned to 12126d8942)