emit_otel()

Create a final action that sends a validation’s metrics to OpenTelemetry.

Usage

emit_otel(
    *,
    meter_name="pointblank",
    enable_metrics=True,
    enable_tracing=False,
    enable_logging=False,
    meter_provider=None,
    tracer_provider=None,
    logger_provider=None,
    metric_prefix="pb.validation",
    log_level="warning",
    extra_attributes=None,
)

The emit_otel() function returns an action for FinalActions that sends validation metrics to OpenTelemetry (OTel) at the end of every interrogate() call. With the action in the validation plan, every run reports to the monitoring backend without any code after the validation. That suits validations that run on a schedule, such as checks in a data pipeline.

The action sends the metrics of OTelExporter, with these differences:

It sends metrics only. The enable_tracing=, enable_logging=, tracer_provider=, logger_provider= and log_level= arguments currently have no effect. Use OTelExporter.export() after the interrogation for traces and log records.
It doesn’t send the step duration histogram ({prefix}.step.duration), so there are 11 metrics instead of 12.
The only attributes are pb.tbl_name (when the validation has a tbl_name=) and those in extra_attributes=.

Without meter_provider=, the metrics go to OTel’s global meter provider, which discards them until the application sets up the OTel SDK. The OTel packages are an optional dependency, installed with pip install pointblank[otel].

Parameters

meter_name: str = "pointblank": The name that identifies Pointblank as the source of the metrics (the OTel instrumentation scope). The default is "pointblank".
enable_metrics: bool = True: Whether to send the metrics. The default is True.
enable_tracing: bool = False: Currently has no effect. The default is False.
enable_logging: bool = False: Currently has no effect. The default is False.
meter_provider: Any | None = None: An OTel MeterProvider for the metrics. The default is None, which uses the global meter provider.
tracer_provider: Any | None = None: Currently has no effect. The default is None.
logger_provider: Any | None = None: Currently has no effect. The default is None.
metric_prefix: str = "pb.validation": The prefix for the names of the metrics. The default is "pb.validation".
log_level: str = "warning": Currently has no effect. The default is "warning".
extra_attributes: dict[str, str] | None = None: A dictionary of attributes to add to every metric, such as the environment or the name of the pipeline. The default is None.

Returns

Callable: A function to use in FinalActions.

Raises

ImportError: If the OTel packages aren’t installed.

Examples

Let’s add the action to a validation of a small table of orders. The OTel SDK’s InMemoryMetricReader keeps the metrics in memory so we can look at them (in production, the meter provider would send them to a collector instead):

import pointblank as pb
import polars as pl
from opentelemetry.sdk.metrics import MeterProvider
from opentelemetry.sdk.metrics.export import InMemoryMetricReader

reader = InMemoryMetricReader()
meter_provider = MeterProvider(metric_readers=[reader])

orders = pl.DataFrame(
    {
        "amount": [29.99, 150.00, -5.00, 75.50, 0.00],
        "customer": ["Alice", "Bob", None, "Dana", "Eve"],
    }
)

validation = (
    pb.Validate(
        data=orders,
        tbl_name="orders",
        thresholds=(0.05, 0.20, 0.40),
        final_actions=pb.FinalActions(
            pb.emit_otel(meter_provider=meter_provider, extra_attributes={"env": "dev"})
        ),
    )
    .col_vals_not_null(columns="customer")
    .col_vals_gt(columns="amount", value=0)
    .interrogate()
)

metrics = reader.get_metrics_data().resource_metrics[0].scope_metrics[0].metrics

for metric in metrics:
    print(metric.name, metric.data.data_points[0].value)

pb.validation.steps.total 2
pb.validation.steps.passed 0
pb.validation.steps.failed 2
pb.validation.test_units.total 10
pb.validation.test_units.passed 7
pb.validation.test_units.failed 3
pb.validation.pass_rate 0.7
pb.validation.duration 0.005109
pb.validation.threshold.warning 2
pb.validation.threshold.error 2
pb.validation.threshold.critical 1

The metrics were sent during interrogate(). Both steps failed, 7 of the 10 test units passed, and the second step exceeded all three thresholds. Each metric has the pb.tbl_name and env attributes:

dict(metrics[0].data.data_points[0].attributes)

{'pb.tbl_name': 'orders', 'env': 'dev'}

Since FinalActions takes any number of actions, the metrics can be sent along with other notifications. Here, a send_slack_notification() action without a webhook URL prints the message it would send:

validation = (
    pb.Validate(
        data=orders,
        tbl_name="orders",
        thresholds=(0.05, 0.20, 0.40),
        final_actions=pb.FinalActions(
            pb.emit_otel(meter_provider=meter_provider),
            pb.send_slack_notification(),
        ),
    )
    .col_vals_not_null(columns="customer")
    .col_vals_gt(columns="amount", value=0)
    .interrogate()
)


=== Dry Run ===
Would send to Slack:
*Data Validation Summary*
• Highest Severity: critical
• Total Steps: 2
• Failed Steps: 2
• Time: 2026-10-29 12:56:14.594144
==============

The actions run in the order they’re given, so the metrics are sent before the Slack message. The message summarizes the same results as the metrics: both steps failed, and the highest threshold level reached was critical.

Parameters

Returns

Raises

Examples

See Also