Setting up training for alert seasonality detection
Set up alert seasonality detection to train policies. Then, choose how to deploy the policies.
For more information about alert seasonality, see Alert seasonality detection algorithm.
Before you begin
To train a definition for alert seasonality detection, you need event data. By default, this AI algorithm can load up to 90 days of historical event data. A set of data creates a set of policies. For more information, see Prerequisites.
This AI algorithm can start to discover seasonal trends after a pattern is learned, which is typically ranging between days or months. For example, if an alert instance occurs on each weekday at two PM, the seasonality algorithm will detect a daily seasonal pattern after one week. However, if this only occurs on Tuesdays at a given time, then it will take multiple weeks to detect this pattern in the alert data.
About this task
There are multiple parts to this task:
- Starting the training setup
- Scheduling the training
- Select data
- Deciding how to deploy the training
- Running the training again
- What to do next
Starting the training setup
Go to the AI Model Management and complete some setup tasks.
- In the Cloud Pak for AIOps home page, click the navigation icon at the upper-left corner of the screen to go to the main navigation menu.
- In the main navigation menu, click Operate > AI model management to open the AI Model Management.
- In the Training tab, click Set up training within the Alert seasonality detection tile under the Trainable AI algorithms section.
Scheduling the training
Decide whether to run the training set up on demand or to schedule it to run on an ongoing basis.
Note: When you schedule each of the trainable AI algorithms, it is strongly recommended to schedule training on quieter periods of the day, and to stagger them out. For example, schedule training runs at least one hour apart from each other. This helps to spread the analytical processing load out.
- Proceed as follows:
- To run the training set up on demand, ensure that Schedule to run is set to Off, and go to step 3.
- To run the training set up on a schedule, set Schedule to run to Yes, and go to the next step.
- Schedule the run. Click this option to specify a schedule. You can specify a start date with an optional end date, a frequency, and a time based on Coordinated Universal Time (UTC).
- Click Next to move to the next panel.
Select data
Select a date range (either Preset or Custom) for the event data that will be included in the AI training. Including additional data typically improves the effectiveness of the AI models and policies that are generated by the training.
Deciding how to deploy the training
The option to review training results before deployment is disabled for this algorithm. However, you can choose when to deploy the policies that are generated by training the algorithm.
- To review and manually deploy the model, set Deployment type to Manual.
- To automatically deploy the model, set Deployment type to On completion.
Click Done to save the training setup for the algorithm.
Running the training again
- If you run the training again with the same data, expect to receive the same policies.
- If you run the training again with different data, you might receive new policies.
What to do next
Now that the training setup is complete, you can train the algorithm. For more details, see Launching the training.