Amazon Lookout for Equipment uses the data from your sensors to detect abnormal equipment behavior, so you can take action before machine failures occur and avoid unplanned downtime. Use the content from this folder to learn how to build a native CloudWatch dashboard that will help you visualize the models and schedulers available in your AWS account.
To get the most out of this content, you need to have at least one Lookout for Equipment model trained in your account and one scheduler configured. Your scheduler must have already run and generated some inference data that you will be able to visualize on a dashboard.
You can check out the Getting Started section of this very repository to train your first model and you can let your inference run for an hour to start generating some data.
Create an AWS account if you do not already have one: then you can setup your environment automatically by deploying the following CloudFormation template.
In another browser window, login to your AWS account. Once you have done that, open one of the link below (depending on the region closest to you) in a new tab to start the process of deploying the items you need via CloudFormation.
Note: This material is designed to work in the regions where the service is available. Using other regions will cause issues.
| Region | CloudFormation Stack | |
|---|---|---|
| US East (N. Virginia) | us-east-1 |  |
| Europe (Ireland) | eu-west-1 | |
| Asia Pacific (Seoul) | ap-northeast-2 | |
After you clicked on the Launch stack button above, you will be brought to
the Create stack screen. Start by clicking Next at the bottom:
On the parameters page, you will have to fill in:
SnapshotBucket: A unique name for the bucket where the dashboard snapshots and temporary files will be collectedSnapshotRuns: The frequency (daily or weekly) at which the inference dashboard snapshot will be taken and sent by emailTargetEmail: The email where the snapshots will be sent
The target email must be registered in Amazon Simple Email Service (SES). To do this:
- Browse to the SES Console
- In the left menu, click on
Email addresses - Then, click on
Verify a New Email Address - Fill in the address you want to use and click
Verify this Email address - You will shortly receive an email in your inbox: click on the
linklink to verify your address.
Once you're fine with these parameters, click Next:
The stack configuration options page will also be fine with the default values,
click Next to proceed to the Review page. Scroll again to the bottom, check
the box to enable the template to create new IAM resources and then click on
Create stack:
CloudFormation will take a few minutes to create the resources described above on your behalf:
Once it is completed, you will see a green CREATE_COMPLETE message next to the name of the stack (on the left panel of the CloudFormation console), indicating that the deployment is completed.
This template will deploy the following component in your AWS account:
Once the template is deployed, navigate to your CloudWatch home page
and click on the Dashboards link on the left menu bar:
This will bring you to a page with the list of all the CloudWatch dashboards you have access to in your account. You should find two new dashboards that you are going to use to visualize your Amazon Lookout for Equipment models and schedulers:
Click on the dashboard with a name starting by Lookout-for-Equipment-Models:
this will open a dashboard summarising all the models that exist in your
account. By default, only the last 50 models trained in the past 3 months are shown:
You can use the date picker on the top right to limit the models shown by their training date. Your models are organized as such:
- The first column will list all the datasets from the current region in your account
- The model names are listed in the second column: if several models were built using the same dataset, the dataset name is not repeated for the sake of clarity
- In the third column, a training status will let you know if this model was successfully trained or not. Only a successful training will give you access to the actions of the fourth column
- The last column gives you two possible actions to perform on trained models:
- You can request the creation of a specific dashboard for a trained model: this will
create a new CloudWatch dashboard specifically for this model and will also configure
and start a
CloudWatch Synthetics
canary that will send you a dashboard screenshot
to the email you configured when you deployed the CloudFormation template. Dashboard
and Synthetics creation can take a few seconds, then the models list dashboard will be
refreshed and the
Viewbutton will be enabled - The
Viewbutton is enabled when a dashboard exists for a given model (theCreate) button will then be disabled. Click on this button to navigate to this model's specific CloudWatch dashboard
- You can request the creation of a specific dashboard for a trained model: this will
create a new CloudWatch dashboard specifically for this model and will also configure
and start a
CloudWatch Synthetics
canary that will send you a dashboard screenshot
to the email you configured when you deployed the CloudFormation template. Dashboard
and Synthetics creation can take a few seconds, then the models list dashboard will be
refreshed and the
When you click on the View button of any trained model, you are brought to another
CloudWatch dashboard dedicated to this model. Several widgets are preconfigured:
- Model details: this widget gives you a reminder of the dataset name, the training period and the evaluation period.
- Detected anomalies: the detected anomalies are displayed for the evaluation period. The dropdown at the top of the widget lets you select which signal you want to plot above the detected anomalies ribbon. This widget also plots the number of daily event detected and the average duration of the detected events over the evaluation period of the model:
- Aggregated signal importance: his horizontal bar chart aggregates the average importance
of up to 15 signals across the evaluation period.
- While the native console from Amazon Lookout for Equipment lets you click on every event detected to plot the associated signal importance, this plot lets you understand if there is a systematic bias towards the same sensors when an event is detected. In the following example, you can see that signal-078 is contributing, in average, a lot more than the other signals to the events detected.
- The black dotted line materialize the value that each signal importance would have if they were all equally contributing to every anomalies detected. The dataset illustrated here had 122 signals. If every signal was contributing the same way, they would each be ranked at approximately 1/122 = 0.82%. We can now say that the higher contribution from signal-078 is definitely statistically significant.
- Every signals higher that the black dotted line will have a red bar, whereas the other ones will have a green bar.
- The last widget plots the evolution of the feature importance of each signal over time. When
numerous events are detected, this plot can become very crowded. At this stage, the feature
importance is aggregated at the daily level to help better understand the dynamics of each
signal.
- During the period ranging from 2015-10-15 to 2015-11-01, you can actually see a different dynamic than during the last period (where a green signal is becoming preponderant)
- This widget will help you understand if it is possible to match the signal contributions relationships to different types of failure.
- A widget initially positionned on the right of this plot contains the associated legend
Click on the dashboard with a name starting by Lookout-for-Equipment-Schedulers:
this will open a dashboard summarising all the schedulers that exist in your
account. By default, only the last 50 schedulers created in the past 3 months are shown:
From this dashboard, you can either Start or Stop the schedulers. Bear in mind that running schedulers incurs costs: do not forget to stop a scheduler that you have no more use of. As with the trained models evaluation, you can also request the creation of specific dashboard dedicated to a scheduler.
When you create a scheduler, you also create a CloudWatch Synthetics
canary that will send you a weekly dashboard screenshot to the email you configured when you
deployed the CloudFormation template. This canary is configured by default to run every
Monday morning at 6am UTC. Dashboard and Synthetics creation can take a few seconds,
then the schedulers list dashboard will be refreshed and the View button will be enabled.
When you click on the View button of an existing inference scheduler, you are brought to
another CloudWatch dashboard dedicated to this scheduler. A couple of widgets are preconfigured:
- Scheduler details: this widget gives you reminder of the scheduler name and how it is configured. It also details what it expects as an input (format and location of the input CSV file to run inference with). This can be useful when you run into issues with a scheduler that doesn't seem to find your fresh data.
- Last execution diagnostics: this widget displays the signal ranking for the last anomaly caught by Amazon Lookout for Equipment:
This folder is structured as followed:
. cloudwatch-dashboard/
|
├── README.md <-- This instruction file
|
├── assets/ <-- Pictures used throughout the README file
|
├── layers/ <-- Lambda layers
| |
| ├── lookout-equipment/ <-- Utility layers used by the Lambda functions
| |
| └── create-layer.sh <-- Script to generate a layer from a public
| package such as `pandas` or `matplotlib`
|
├── lambdas/ <-- Lambda functions source code
|
└── synthetics/ <-- Synthetics canary code to take dashboard snapshots
Please contact Michaël HOARAU or raise an issue on this repository.
See CONTRIBUTING for more information.
This source code is licensed under the MIT-0 License. See the LICENSE file.