Designing and Implementing a Data Science Solution on Azure: DP-100 Exam | Question 1
Question
You are working for an agricultural company which monitors its land areas using an IoT solution, including several types of environmental sensors, local weather stations, drones etc.
These data sources gather field data continuously and send it to a central location where processing has to be done once a day (preferably in night hours) and predictions have to be generated for the following day.
Your task is to build an ML solution to ingest the daily batches of data regularly, pre-process and feed it to an inference service.
You decide to use Azure ML pipelines and run them at predefined intervals.
Set schedule recurrence Create schedule Enable schedule Publish pipeline Retrieve pipeline id Which steps should you use in your script in what order?
Answers
Explanations
Click on the arrows to vote for the correct answer
A. B. C. D.Answer: B.
Option A is incorrect because the schedule must be submitted in order to become active.
Once it is submitted, it becomes active automatically.
Enabling/disabling have effect only on existing schedules.
Option B is CORRECT because so that you can initiate the scheduled runs of a pipeline, the pipeline must be created and published first.
Then you have to create a ScheduleRecurrence (trigger) object and, finally the Schedule itself, connecting the pipeline (via its id) and the trigger object must be created and run.
You can watch the scheduled pipelines in the Endpoints section of the ML Studio.
Option C is incorrect because the Schedule object connects the pipeline and a triggering event together.
Therefore, the defining the schedule recurrence (i.e.
the trigger) must precede submitting the schedule.
Option D is incorrect because the scheduled pipeline gets into ‘Active' status automatically.
There is no need enabling them explicitly.
Reference - Monitor scheduled pipelines.
/answer/imgycZhT1mnlQfDG1x8Mv0O6BXyfij48oYiVz2I8XpVHKnJnK14q_aoeS5zi6G9bYat9n8-7sGWC4iUbPeGeeeJ7ZjaHC-OWd-Q-2CVK4Rxmif50REcF7RfZPEEtx7q055hBZL_cpQQKwl1k1u9AQ)
Reference:
To build an ML solution for the agricultural company, you will need to use Azure ML pipelines. You will need to create a pipeline to ingest the daily batches of data regularly, pre-process it, and feed it to an inference service. The pipeline should also be scheduled to run at predefined intervals, preferably during night hours.
The steps required to create a scheduled pipeline are as follows:
Publish pipeline: Publish the pipeline to make it available for execution. This is done using the Azure Machine Learning SDK or Azure Portal.
Set schedule recurrence: Set the recurrence pattern for the pipeline to define when and how frequently the pipeline should run. This can be done using the Azure Machine Learning SDK or Azure Portal.
Enable schedule: Enable the pipeline schedule to start executing the pipeline at the defined recurrence. This can be done using the Azure Machine Learning SDK or Azure Portal.
Retrieve pipeline ID: Retrieve the pipeline ID that was generated when the pipeline was published. This ID is required to trigger the pipeline execution.
Therefore, the correct order of steps to create a scheduled pipeline using Azure ML pipelines for the agricultural company's IoT solution is option D: 4, 5, 2, 1, 3.
First, you need to publish the pipeline (step 4) to make it available for execution. Then, you need to retrieve the pipeline ID (step 5) to trigger the pipeline execution. After that, you can set the schedule recurrence (step 2) to define when and how frequently the pipeline should run. Finally, you need to enable the schedule (step 1) to start executing the pipeline at the defined recurrence.