Sensors#

Sensors allow you to instigate runs based on any external state change.

Relevant APIs#

Overview#

Sensors are definitions in Dagster that allow you to instigate runs based on some external state change automatically. For example, you can:

• Launch a run whenever a file appears in an s3 bucket
• Launch a run whenever another pipeline materializes a specific asset
• Launch a run whenever an external system is down

Sensors have several important properties:

• Each sensor targets a specific pipeline
• A sensor optionally defines tags, a mode, and a solid selection for the targeted pipeline.
• A sensor defines an evaluation function that returns either:
  • One or more RunRequest objects. Each run request launches a run.
  • An optional SkipReason, which specifies a message which describes why no runs were requested.

The Dagster Daemon runs each sensor evaluation function on a tight loop. If you are using sensors, make sure to follow the instructions on the DagsterDaemon page to run your sensors.

Defining a sensor#

To define a sensor, use the @sensor decorator. The decorated function is called the execution_fn and must have context as the first argument. The context is a SensorExecutionContext

Let's say you have a pipeline that logs a filename that is specified in the solid configuration of the process_file solid:

from dagster import solid, pipeline


@solid(config_schema={"filename": str})
def process_file(context):
    filename = context.solid_config["filename"]
    context.log.info(filename)


@pipeline
def log_file_pipeline():
    process_file()

You can write a sensor that watches for new files in a specific directory and yields a RunRequest for each new file in the directory. By default, this sensor every 30 seconds.

import os
from dagster import sensor, RunRequest


@sensor(pipeline_name="log_file_pipeline")
def my_directory_sensor(_context):
    for filename in os.listdir(MY_DIRECTORY):
        filepath = os.path.join(MY_DIRECTORY, filename)
        if os.path.isfile(filepath):
            yield RunRequest(
                run_key=filename,
                run_config={"solids": {"process_file": {"config": {"filename": filename}}}},
            )

This sensor iterates through all the files in MY_DIRECTORY and yields a RunRequest for each file.

Once my_directory_sensor is added to a repository with log_file_pipeline, it can be enabled and used.

Idempotence and Cursors#

When instigating runs based on external events, you usually want to run exactly one pipeline run for each event. There are two ways to define your sensors to avoid creating duplicate runs for your events: using run_key and using a cursor.

Idempotence using run keys#

In the example sensor above, the RunRequest is constructed with a run_key.

yield RunRequest(
        run_key=filename,
        run_config={"solids": {"process_file": {"config": {"filename": filename}}}},
    )

Dagster guarantees that for a given sensor, at most one run is created for each RunRequest with a unique run_key. If a sensor yields a new run request with a previously used run_key, Dagster skips processing the new run request.

In the example, a RunRequest is requested for each file during every sensor evaluation. Therefore, for a given sensor evaluation, there already exists a RunRequest with a run_key for any file that existed during the previous sensor evaluation. Dagster skips processing duplicate run requests, so Dagster launches runs for only the files added since the last sensor evaluation. The result is exactly one run per file.

Run keys allow you to write sensor evaluation functions that declaratively describe what pipeline runs should exist, and helps you avoid the need for more complex logic that manages state. However, when dealing with high-volume external events, some state-tracking optimizations might be necessary.

Sensor optimizations using cursors#

When writing a sensor that deals with high-volume events, it might not be feasible to yield a RunRequest during every sensor evaluation. For example, you may have an s3 storage bucket that contains thousands of files.

When writing a sensor for such event sources, you can maintain a cursor that limits the number of yielded run requests for previously processed events. The sensor context, provided to every sensor evaluation function, has a cursor property and a update_cursor method for sensors to track state across evaluations.

• cursor: A cursor field on SensorExecutionContext that returns the last persisted cursor value from a previous evaluation.
• update_cursor: A method on SensorExecutionContext that takes a string to persist and make available to future evaluations.

Here is a somewhat contrived example of our directory file sensor using a cursor for updated files.

@sensor(pipeline_name="log_file_pipeline")
def my_directory_sensor_cursor(context):
    last_mtime = float(context.cursor) if context.cursor else 0

    max_mtime = last_mtime
    for filename in os.listdir(MY_DIRECTORY):
        filepath = os.path.join(MY_DIRECTORY, filename)
        if os.path.isfile(filepath):
            fstats = os.stat(filepath)
            file_mtime = fstats.st_mtime
            if file_mtime <= last_mtime:
                continue

            # the run key should include mtime if we want to kick off new runs based on file modifications
            run_key = f"{filename}:{str(file_mtime)}"
            run_config = {"solids": {"process_file": {"config": {"filename": filename}}}}
            yield RunRequest(run_key=run_key, run_config=run_config)
            max_mtime = max(max_mtime, file_mtime)

    context.update_cursor(str(max_mtime))

For sensors that consume multiple event streams, you may need to serialize and deserialize a more complex data structure in and out of the cursor string to keep track of the sensor's progress over the multiple streams.

Evaluation Interval#

By default, the Dagster Daemon runs a sensor 30 seconds after the previous sensor evaluation finishes executing. You can configure the interval using the minimum_interval_seconds argument on the @sensor decorator.

It's important to note that this interval represents a minimum interval between runs of the sensor and not the exact frequency the sensor runs. If you have a sensor that takes 2 minutes to complete, but the minimum_interval_seconds is 5 seconds, the fastest Dagster Daemon will run the sensor is every 2 minutes and 5 seconds. The minimum_interval_seconds only guarantees that the sensor is not evaluated more frequently than the given interval.

For example, here are two sensors that specify two different minimum intervals:

@sensor(pipeline_name="my_pipeline", minimum_interval_seconds=30)
def sensor_A(_context):
    yield RunRequest(run_key=None, run_config={})


@sensor(pipeline_name="my_pipeline", minimum_interval_seconds=45)
def sensor_B(_context):
    yield RunRequest(run_key=None, run_config={})

These sensor definitions are short, so they run in less than a second. Therefore, you can expect these sensors to run consistently around every 30 and 45 seconds, respectively.

Skipping sensor evaluations#

For debugging purposes, it is often useful to describe why a sensor might not yield any runs for a given evaluation. The sensor evaluation function can yield a SkipReason with a string description that will be displayed in Dagit.

For example, here is our directory sensor that now provides a SkipReason when no files are encountered:

@sensor(pipeline_name="log_file_pipeline")
def my_directory_sensor_with_skip_reasons(_context):
    has_files = False
    for filename in os.listdir(MY_DIRECTORY):
        filepath = os.path.join(MY_DIRECTORY, filename)
        if os.path.isfile(filepath):
            yield RunRequest(
                run_key=filename,
                run_config={"solids": {"process_file": {"config": {"filename": filename}}}},
            )
            has_files = True
    if not has_files:
        yield SkipReason(f"No files found in {MY_DIRECTORY}.")

Testing sensors#

To quickly preview what an existing sensor would generate when evaluated, you can run the CLI command dagster sensor preview my_sensor_name.

In order to unit test sensors, you can invoke the sensor directly. This will return all the run requests yielded by the sensor. The config obtained from these can be validated using the validate_run_config function.

from dagster import validate_run_config


@sensor(pipeline_name="log_file_pipeline")
def sensor_to_test(_context):
    yield RunRequest(
        run_key="foo",
        run_config={"solids": {"process_file": {"config": {"filename": "foo"}}}},
    )


def test_sensor():
    for run_request in sensor_to_test(None):
        assert validate_run_config(log_file_pipeline, run_request.run_config)

Notice that since we did not use the context argument in our sensor, we don't have to provide a context object. However, if we do in fact need the context object for our sensor, we can provide it via build_sensor_context. Consider again the my_directory_sensor_cursor example.

@sensor(pipeline_name="log_file_pipeline")
def my_directory_sensor_cursor(context):
    last_mtime = float(context.cursor) if context.cursor else 0

    max_mtime = last_mtime
    for filename in os.listdir(MY_DIRECTORY):
        filepath = os.path.join(MY_DIRECTORY, filename)
        if os.path.isfile(filepath):
            fstats = os.stat(filepath)
            file_mtime = fstats.st_mtime
            if file_mtime <= last_mtime:
                continue

            # the run key should include mtime if we want to kick off new runs based on file modifications
            run_key = f"{filename}:{str(file_mtime)}"
            run_config = {"solids": {"process_file": {"config": {"filename": filename}}}}
            yield RunRequest(run_key=run_key, run_config=run_config)
            max_mtime = max(max_mtime, file_mtime)

    context.update_cursor(str(max_mtime))

This sensor makes use of the context argument, and thus to invoke it, we need to provide one.

from dagster import build_sensor_context


def test_my_directory_sensor_cursor():
    context = build_sensor_context(cursor="0")
    for run_request in my_directory_sensor_cursor(context):
        assert validate_run_config(log_file_pipeline, run_request.run_config)

Monitoring sensors in Dagit#

You can monitor and operate sensors in Dagit. There are multiple views that help with observing sensor evaluations, skip reasons, and errors.

To view the sensors page, click the "All sensors" in the left-hand navigation pane. Here you can turn sensors on and off using the toggle.

If you click on any sensor, you can monitor all sensor evaluations and runs created:

If your sensor throws an error or yields a skip reason, the sensor timeline view will display more information about the errors and skips:

Examples#

Asset sensors#

A useful pattern is to create a sensor that checks for new AssetMaterialization events for a particular asset key. This can be used to kick off a pipeline that computes downstream assets or notifies appropriate stakeholders.

One benefit of this pattern is that it enables cross-pipeline and even cross-repository dependencies. Each pipeline run instigated by an asset sensor is agnostic to the pipeline event that caused it.

Here is an example of a sensor that listens for asset materializations for a given asset key my_table:

from dagster import AssetKey


@sensor(pipeline_name="my_pipeline")
def my_asset_sensor(context):
    events = context.instance.events_for_asset_key(
        AssetKey("my_table"), after_cursor=context.cursor, ascending=False, limit=1
    )
    if not events:
        return

    record_id, event = events[0]  # take the most recent materialization
    yield RunRequest(
        run_key=str(record_id), run_config={}, tags={"source_pipeline": event.pipeline_name}
    )
    context.update_cursor(str(record_id))

Pipeline failure sensor#

If you want to act on pipeline failures, e.g., you need to send an alert to a monitoring service on pipeline failure. You can write a sensor that monitors Dagster's runs table and launches a specialized "alert" pipeline for each failed run.

For example, you can write an "alert" pipeline that sends a slack message when it runs. Note that the pipeline depends on a slack resource:

@solid(required_resource_keys={"slack"})
def slack_message_on_failure_solid(context):
    message = f"Solid {context.solid.name} failed"
    context.resources.slack.chat.post_message(channel="#foo", text=message)


@pipeline(
    mode_defs=[
        ModeDefinition(name="test", resource_defs={"slack": ResourceDefinition.mock_resource()}),
        ModeDefinition(name="prod", resource_defs={"slack": slack_resource}),
    ]
)
def failure_alert_pipeline():
    slack_message_on_failure_solid()

Then, you can define a sensor that fetches the failed runs from the runs table via context.instance, and instigates a failure_alert_pipeline run for every failed run. Note that we use the failed run's id as the run_key to prevent sending an alert twice for the same pipeline run.

@sensor(pipeline_name="failure_alert_pipeline", mode="prod")
def pipeline_failure_sensor(context):
    runs = context.instance.get_runs(
        filters=PipelineRunsFilter(
            pipeline_name="your_pipeline_name",
            statuses=[PipelineRunStatus.FAILURE],
        ),
    )
    for run in runs:
        # Use the id of the failed run as run_key to avoid duplicate alerts.
        yield RunRequest(run_key=str(run.run_id))

If you would like to set up success or failure handling policies on solids, you can find more information on the Solid Hooks page.

S3 sensors#

For pipelines that should initiate new runs for new paths in an s3 bucket, the dagster-aws package provides the useful helper function get_s3_keys.

Here is an example of a sensor that listens to a particular s3 bucket my_s3_bucket:

from dagster_aws.s3.sensor import get_s3_keys


@sensor(pipeline_name="my_pipeline")
def my_s3_sensor(context):
    new_s3_keys = get_s3_keys("my_s3_bucket", since_key=context.last_run_key)
    if not new_s3_keys:
        yield SkipReason("No new s3 files found for bucket my_s3_bucket.")
        return
    for s3_key in new_s3_keys:
        yield RunRequest(run_key=s3_key, run_config={})