I am facing this issue :

Exception: Error in sensor publish_on_success_sensor: Sensor returned a RunRequest with job_name job2. Expected one of: ['job1_name']

Here is the code that must be the issue :

@run_status_sensor(
    run_status=DagsterRunStatus.SUCCESS,
    request_job=job1,
    default_status=DefaultSensorStatus.RUNNING,
)
def publish_on_success_sensor(context):
    if context.dagster_run.job_name == job1.name:
        yield RunRequest(run_key=None, run_config={}, job_name="job2")
    else:
        return SkipReason("Not launching on jobs other than job1")

What can be the root cause ?

The error message you're encountering indicates that the

publish_on_success_sensor

sensor is returning a

RunRequest

for a job named

job2

, but the sensor is expected to return a

RunRequest

for

job1

or a list of jobs that includes

job1_name

. Based on the code snippet you provided, the

publish_on_success_sensor

is decorated with

@run_status_sensor

and specifies

request_job=job1

. This means that the sensor is configured to request runs for

job1

when it yields a

RunRequest

. However, inside the sensor function, you are attempting to yield a

RunRequest

for a job named

job2

, which is not the job specified in the

request_job

parameter of the sensor decorator. To resolve this issue, you need to ensure that the

RunRequest

you yield is for the job specified in the

request_job

parameter of the sensor decorator. If you want to trigger

job2

from this sensor, you need to include

job2

in the list of jobs that the sensor is allowed to request runs for. This can be done by using the

request_jobs

parameter (if available and supported by your version of Dagster) to specify multiple jobs that the sensor can trigger. Here's an example of how you might modify the sensor to include multiple jobs:

python @run_status_sensor(     run_status=DagsterRunStatus.SUCCESS,     request_jobs=[job1, job2],  # Assuming request_jobs is supported and allows multiple jobs     default_status=DefaultSensorStatus.RUNNING, ) def publish_on_success_sensor(context):     if context.dagster_run.job_name == job1.name:         yield RunRequest(run_key=None, run_config={}, job_name="job2")     else:         return SkipReason("Not launching on jobs other than job1")

Please note that the

request_jobs

parameter is marked as experimental in the documentation and may not be available in all versions of Dagster. If your version of Dagster does not support

request_jobs

, you may need to define separate sensors for each job you want to trigger or update your sensor logic to only yield

RunRequest

for

job1

. If you continue to face issues, please ensure that your Dagster version supports the features you are trying to use and that you are following the latest documentation for run status sensors.

@Scout if I take the code snippet you gave me, will my job still be triggered on job1's success ?

The code snippet I provided was intended to illustrate how you might configure a sensor to trigger multiple jobs. However, the

request_jobs

parameter I used in the example is not actually part of the Dagster API based on the documentation available to me. I apologize for the confusion. In Dagster, a

run_status_sensor

is typically configured to monitor the status of one or more jobs and can trigger another job upon a successful run. However, the sensor must be configured correctly to do so. To trigger

job2

upon the successful completion of

job1

, you would need to set up your sensor to monitor

job1

and then yield a

RunRequest

for

job2

. Here's how you might write a sensor that does this:

python @run_status_sensor(     run_status=DagsterRunStatus.SUCCESS,     request_job=job2,  # This should be the job you want to trigger     default_status=DefaultSensorStatus.RUNNING, ) def publish_on_success_sensor(context):     if context.dagster_run.job_name == job1.name:         # Trigger job2 if job1 succeeds         yield RunRequest(run_key=None, run_config={}, job_name=job2.name)     else:         # Skip if the job that succeeded is not job1         return SkipReason(f"Not launching job2 because the successful job was not {job1.name}")

In this revised example, the sensor is configured with

request_job=job2

, which means it is set up to trigger

job2

. Inside the sensor function, it checks if the successful job is

job1

and, if so, yields a

RunRequest

for

job2

. Please ensure that

job1

and

job2

are defined correctly in your code and that

job1.name

and

job2.name

match the actual names of the jobs you are working with. Remember to consult the latest Dagster documentation for the correct usage of sensors and any version-specific features or changes.