apache/beam · error · RuntimeError

f'Failed to create MonitoringInfo for urn

Error message

f'Failed to create MonitoringInfo for urn {urn} type {type_urn} labels {labels} and payload {payload}'

What it means

create_monitoring_info builds a metrics_pb2.MonitoringInfo protobuf. If the arguments (urn, type_urn, labels, payload) have types the protobuf constructor rejects, it raises TypeError, which is re-raised as a RuntimeError with the full argument values included for debugging. The library does this so callers get a descriptive message instead of a bare protobuf TypeError.

Solutions

Ensure all label keys and values are str (cast with str() or decode bytes) before calling the metric helpers.
Ensure payload is bytes (payload.encode() for str, or the serialized proto bytes).
Inspect the chained exception ('from e') message from protobuf to see exactly which field type was rejected.
Ensure labels is a dict or None, not a list of tuples; use dict(labels) if converting.

Example fix

# before
metrics = Metrics.get_namespace(step).counter(label_id, labels={"step": 3})
# after
metrics = Metrics.get_namespace(step).counter(label_id, labels={"step": "3"})

Defensive patterns

Strategy: try-catch

Validate before calling

assert isinstance(labels, (dict, type(None))) and all(isinstance(k, str) and isinstance(v, str) for k, v in (labels or {}).items())
assert isinstance(payload, (bytes, bytearray, str, type(None)))

Type guard

def valid_labels(labels): return labels is None or (isinstance(labels, dict) and all(isinstance(k, str) and isinstance(v, str) for k, v in labels.items()))

Try / catch

try:
    mi = create_monitoring_info(urn, type_urn, labels, payload)
except RuntimeError as e:
    logger.error("bad monitoring info args: %s", e); labels = {k: str(v) for k, v in labels.items()}; mi = create_monitoring_info(urn, type_urn, labels, payload)

Prevention

Coerce all label values to str at metric-creation sites
Serialize payloads to bytes explicitly
Wrap dynamic label construction in a helper that validates types

When it happens

Trigger: Calling int64_counter/int64_distribution/int64_gauge (or their _user_ variants) with non-string labels keys/values, labels that are not a dict, a payload that is not bytes/str, or passing a labels dict containing non-primitive values into create_monitoring_info.

Common situations: Metric labels built from dynamic data (e.g. f-strings with None, ints, or bytes instead of str); a payload passed as a dict instead of a serialized bytes string; SDK-internal callers after a Beam version changed the MonitoringInfo proto field types.

Understand the failure class

Background: "Must be a positive integer", "Invalid value", "Unsupported": the invalid-argument-value error family, when a library rejects the value you pass — this error's family across 35 libraries.

Related errors

AI-assisted analysis of apache/beam@12126d8942 (2026-09-13). Data as JSON: /api/errors/a568a48244a9f9c5. Report an issue: GitHub.

Appendix: source

Thrown at sdks/python/apache_beam/metrics/monitoring_infos.py:399

      labels)


def create_monitoring_info(
    urn, type_urn, payload, labels=None) -> metrics_pb2.MonitoringInfo:
  """Return the monitoring info for the URN, type, metric and labels.

  Args:
    urn: The URN of the monitoring info/metric.
    type_urn: The URN of the type of the monitoring info/metric.
        i.e. beam:metrics:sum_int_64, beam:metrics:latest_int_64.
    payload: The payload field to use in the monitoring info.
    labels: The label dictionary to use in the MonitoringInfo.
  """
  try:
    return metrics_pb2.MonitoringInfo(
        urn=urn, type=type_urn, labels=labels or {}, payload=payload)
  except TypeError as e:
    raise RuntimeError(
        f'Failed to create MonitoringInfo for urn {urn} type {type_urn} '
        f'labels {labels} and payload {payload}') from e


def is_counter(monitoring_info_proto):
  """Returns true if the monitoring info is a coutner metric."""
  return monitoring_info_proto.type in COUNTER_TYPES


def is_gauge(monitoring_info_proto):
  """Returns true if the monitoring info is a gauge metric."""
  return monitoring_info_proto.type in GAUGE_TYPES


def is_distribution(monitoring_info_proto):
  """Returns true if the monitoring info is a distrbution metric."""
  return monitoring_info_proto.type in DISTRIBUTION_TYPES

View on GitHub (pinned to 12126d8942)