Create a custom rule
Create custom rules by describing them in plain language. The AI rule builder turns your request into a validated rule that checks your FSx for ONTAP resources.
-
You have configured AWS credentials with read-only access.
-
You have ONTAP credentials with viewer permissions.
-
You have a link or agent configured for FSx for ONTAP connectivity.
-
You understand that custom rule creation uses NetApp's Bedrock AI service.
The rule creation process walks you through the steps: defining the rule, reviewing how the AI interpreted it, and checking the results before scheduling regular evaluations. You stay in control at every step. The system suggests a rule based on your prompt, but nothing moves forward until you confirm it.
Create a rule
-
Log in using one of the console experiences.
-
Select the menu
and then select Well-architected. -
From the Well-architected menu, select Rule builder.
When you open the rule builder, it shows sample prompts to help you write your custom rule. You can select Copy prompt to copy a sample prompt into the AI dialog text box.
-
In the Create a rule field, enter your rule description in natural language. Be as specific as possible about:
-
What resource type you're checking (file system, volume, or cache).
-
What condition you're validating.
-
What threshold defines compliance.
-
Why this rule matters (optional but recommended).
-
-
Select
.The AI analyzes your prompt and generates a rule definition.
Review the AI interpretation
The rule builder displays:
-
Parsed interpretation: How the AI understood your prompt in plain language.
-
Rule definition: The structured rule including resource type, condition logic, and evaluation criteria.
-
Reasoning: Why the AI interpreted your prompt this way.
-
Confidence indicator: How confident the system is in the interpretation.
-
Review the interpretation to make sure it matches what you intended.
Check that the rule definition is correct.
If the interpretation is unclear or wrong, the system may ask questions such as:
* Which resource should be checked?
* What value or limit defines compliance?
* How often should the rule run? -
Respond to the AI's clarifying questions.
The AI might ask you to choose from several options to better understand what you need. Answer each question and select Continue, or select Skip to move past optional questions.
-
After answering all questions, select Continue to create your rule.
Test the rule
Before scheduling the rule, run a one-time test to see which resources it will affect.
-
Select Validate rule.
The system evaluates all resources in scope against your rule logic.
-
Review the validation results:
-
Resources evaluated: Total number of resources checked
-
Compliant: Resources that meet the rule criteria
-
Non-compliant: Resources that do not meet the criteria
-
Failed evaluations: Resources that could not be evaluated (with error details)
-
-
If some resources failed evaluation, review the error reasons:
-
AWS credential errors: Read-only permissions may be insufficient for certain resources
-
Connectivity errors: Link or agent might not be configured for specific file systems
-
ONTAP permission errors: The viewer user might not have access to certain resources
-
Save or schedule the rule
After you test your rule, you can save it as a draft or set it to run regularly.
-
Do one of the following:
-
To schedule regular evaluation, set the evaluation frequency (minimum 1 hour, default 24 hours) and select Schedule. The rule is enabled and runs on the defined schedule.
-
To save without scheduling, select Save as draft. The rule is saved but not evaluated until you schedule it.
-
To discard the rule, select Discard to cancel rule creation and start over.
-
Your custom rule has been created and is listed in the rule catalog. If you set a schedule, it will check your FSx for ONTAP resources at the specified frequency.
Troubleshooting
- No file systems available
-
Set up your AWS credentials and make sure your environment has at least one FSx for ONTAP file system. The rule builder needs a file system to apply rules.
- Low confidence interpretation
-
More information is needed. Answer the questions or rewrite your prompt with clear details about the resource type, limits, and conditions.
- Partial validation failures
-
Check which resources failed and why. Common causes are missing link or agent settings, or not having enough permissions. Fix the problem, then validate again.
- Ambiguous prompt
-
Be more specific about what you want to check. Include the resource type, the condition, and the threshold. See the prompt examples for help.
- Rule targets multiple resource types
-
Create a separate rule for each resource type. Only one resource type per rule is supported.