PAI-FeatureStore FeatureDB Usage Query and Analysis

Architecture: PAI-FeatureStore Instance → FeatureDB Datasource → Feature Views → Usage Statistics

Key Features

1. Total Usage Analysis: Query daily read/write totals across a date range (up to 30 days)
2. Feature View Details: Query specific feature view usage on a given date with sorting by read/write count
3. Project-Level Analysis: Query daily read/write totals for a specific project (up to 30 days)
4. Feature View Trends: Query daily read/write trends for a specific feature view (up to 30 days)
5. Cost Calculation: Calculate costs based on official pricing tiers for different regions

Supported Regions

cn-beijing, cn-hangzhou, cn-shanghai, cn-shenzhen, cn-hongkong, ap-southeast-1, ap-southeast-5, us-west-1, eu-central-1

Installation

Pre-check: Aliyun CLI >= 3.3.3 required

Run aliyun version to verify >= 3.3.3. If not installed or version too low, run curl -fsSL https://aliyuncli.alicdn.com/setup.sh | bash to update, or see references/cli-installation-guide.md for installation instructions.

Then [MUST] run aliyun configure set --auto-plugin-install true to enable automatic plugin installation. [MUST] run aliyun plugin update to ensure that any existing plugins on your local machine are always up-to-date.

[MUST] CLI User-Agent — Every aliyun CLI command invocation must include: --user-agent AlibabaCloud-Agent-Skills/alibabacloud-pai-feature-store-featuredb-usage-query

The PAI-FeatureStore CLI plugin will be automatically installed when first used if auto-plugin-install is enabled.

Timeout: Default 10s is sufficient for all API calls.

Authentication

Pre-check: Alibaba Cloud Credentials Required

Security Rules:

• NEVER read, echo, or print AK/SK values (e.g., echo $ALIBABA_CLOUD_ACCESS_KEY_ID is FORBIDDEN)
• NEVER ask the user to input AK/SK directly in the conversation or command line
• NEVER use aliyun configure set with literal credential values
• NEVER output any string resembling a key (e.g., LTAI5t***, AccessKeyId: xxx, ak-***)
• ONLY use aliyun configure list to check credential status

aliyun configure list

Check the output for a valid profile (AK, STS, or OAuth identity).

If no valid profile exists, STOP here.

Required Parameters

IMPORTANT: Parameter Confirmation — Before executing any command or API call, ALL user-customizable parameters (e.g., RegionId, instance names, CIDR blocks, passwords, domain names, resource specifications, etc.) MUST be confirmed with the user. Do NOT assume or use default values without explicit user approval.

Note: Either WorkspaceId or DatasourceId must be provided. If neither is provided, the skill will list all available FeatureDB datasources for user selection.

RAM Permissions

This skill requires the following RAM permissions. See references/ram-policies.md for detailed policy configuration.

Required Actions:

• paifeaturestore:ListInstances
• paifeaturestore:GetDatasource
• paifeaturestore:ListDatasources
• paifeaturestore:ListDatasourceFeatureViews

Core Workflow

Functional Scope (MUST CHECK FIRST)

This skill ONLY supports querying FeatureDB read/write usage statistics. Any requests outside of "querying historical read/write counts or costs" are out of scope and must be politely declined with a list of supported features.

[MUST] If request is out of scope, reject politely and list supported features:

"Sorry, this Skill only supports querying FeatureDB read/write usage statistics, not [user request]. Supported features: 1) Daily read/write totals 2) Feature view details 3) Usage trends 4) Cost calculation"

Feature Selection Logic

Which features to execute depends on user's request:

Supported Functions

1. Query daily FeatureDB read/write totals from StartDate to EndDate with analysis (max 30-day span)
2. Query per-feature-view read/write counts for a specific date, with sorting and analysis
3. Query daily FeatureDB read/write totals for a specific project from StartDate to EndDate with analysis (max 30-day span)
4. Query daily FeatureDB read/write trends for a specific feature view from StartDate to EndDate with analysis (max 30-day span)
5. Calculate costs based on official pricing and query results

User-Friendly Prompts

Use descriptive names instead of function numbers when prompting, because "Function 3" is meaningless to users who haven't read this document.

Execution Steps

Step 1: Verify PAI-FeatureStore Instance

aliyun paifeaturestore list-instances \
  --region <RegionId> \
  --status Running \
  --user-agent AlibabaCloud-Agent-Skills/alibabacloud-pai-feature-store-featuredb-usage-query

Expected Output: A list of running instances with their InstanceIds.

Error Handling (if no running instances found — STOP, do NOT guess InstanceId):

No running PAI-FeatureStore instances found in this region. Please verify:

1. Is the RegionId correct?
2. Do you need to switch to a different region?
3. To activate PAI-FeatureStore, visit the console

Step 2: Identify FeatureDB Datasource

Case A: User provided WorkspaceId

aliyun paifeaturestore list-datasources \
  --region <RegionId> \
  --instance-id <InstanceId> \
  --type FeatureDB \
  --workspace-id <WorkspaceId> \
  --user-agent AlibabaCloud-Agent-Skills/alibabacloud-pai-feature-store-featuredb-usage-query

Case B: User provided DatasourceId

[MUST] When user provides DatasourceId, you MUST call GetDatasource or ListDatasources to verify the datasource type is FeatureDB. Do NOT skip this validation step.

aliyun paifeaturestore get-datasource \
  --region <RegionId> \
  --instance-id <InstanceId> \
  --datasource-id <DatasourceId> \
  --user-agent AlibabaCloud-Agent-Skills/alibabacloud-pai-feature-store-featuredb-usage-query

Validation: Check the Type field in the response. If Type is NOT FeatureDB, stop and inform the user that this datasource is not a FeatureDB datasource.

Case C: Neither WorkspaceId nor DatasourceId provided

aliyun paifeaturestore list-datasources \
  --region <RegionId> \
  --instance-id <InstanceId> \
  --type FeatureDB \
  --user-agent AlibabaCloud-Agent-Skills/alibabacloud-pai-feature-store-featuredb-usage-query

Note: This API is paginated. Calculate total pages from TotalCount and PageSize, then fetch all pages.

Example pagination logic:

# First call to get TotalCount
FIRST_RESPONSE=$(aliyun paifeaturestore list-datasources \
  --region <RegionId> \
  --instance-id <InstanceId> \
  --type FeatureDB \
  --page-number 1 \
  --page-size 10 \
  --user-agent AlibabaCloud-Agent-Skills/alibabacloud-pai-feature-store-featuredb-usage-query)

# Extract TotalCount and calculate total pages
# Then iterate through all pages
for page in {1..total_pages}; do
  aliyun paifeaturestore list-datasources \
    --region <RegionId> \
    --instance-id <InstanceId> \
    --type FeatureDB \
    --page-number $page \
    --page-size 10 \
    --user-agent AlibabaCloud-Agent-Skills/alibabacloud-pai-feature-store-featuredb-usage-query
done

Present all FeatureDB datasources to the user. If only ONE datasource exists, use it directly.

When MULTIPLE datasources exist, behavior depends on query type:

Error Handling:

• If WorkspaceId or DatasourceId is invalid, fall back to Case C and list all available datasources.

Step 3: Query Usage Statistics

Date Validation (before API call):

• Format: Must be yyyy-MM-dd (e.g., 2024-03-15). If user provides other formats (e.g., 3/15, 2024.03.15), convert to correct format.
• Range: EndDate - StartDate must be ≤ 30 days. If exceeded, ask user to narrow the range or split into multiple queries.
• Boundary: EndDate ≤ today; StartDate ≤ EndDate.

Relative Date Ranges ("last 7 days", "last 30 days"): Calculate StartDate = N days ago, EndDate = yesterday. Run date command separately first, then use the literal result in --start-date/--end-date. NEVER embed $(...) in CLI commands.

Execute queries based on user requirements:

Function 1: Query Total Daily Read/Write Usage

Query daily totals from StartDate to EndDate (default: last 30 days).

aliyun paifeaturestore list-datasource-feature-views \
  --region <RegionId> \
  --instance-id <InstanceId> \
  --datasource-id <DatasourceId> \
  --start-date <StartDate> \
  --end-date <EndDate> \
  --verbose false \
  --show-storage-usage false \
  --page-size 1 \
  --user-agent AlibabaCloud-Agent-Skills/alibabacloud-pai-feature-store-featuredb-usage-query

Output: Extract TotalUsageStatistics from the response.

Function 2: Query Per-Feature-View Usage for a Specific Date

Query all feature views' read/write counts on a specific date, with sorting.

aliyun paifeaturestore list-datasource-feature-views \
  --region <RegionId> \
  --instance-id <InstanceId> \
  --datasource-id <DatasourceId> \
  --start-date <TargetDate> \
  --end-date <TargetDate> \
  --verbose true \
  --show-storage-usage false \
  --all true \
  --sort-by <ReadCount|WriteCount> \
  --order <ASC|DESC> \
  --user-agent AlibabaCloud-Agent-Skills/alibabacloud-pai-feature-store-featuredb-usage-query

Output: Extract FeatureViews array from the response.

Function 3: Query Daily Usage for a Specific Project

Query daily totals for a specific project from StartDate to EndDate (default: last 7 days).

aliyun paifeaturestore list-datasource-feature-views \
  --region <RegionId> \
  --instance-id <InstanceId> \
  --datasource-id <DatasourceId> \
  --start-date <StartDate> \
  --end-date <EndDate> \
  --verbose false \
  --show-storage-usage false \
  --page-size 1 \
  --project-name <ProjectName> \
  --user-agent AlibabaCloud-Agent-Skills/alibabacloud-pai-feature-store-featuredb-usage-query

Output: Extract TotalUsageStatistics from the response.

Error Handling:

• If the API returns an error indicating the project does not exist, recommend querying Function 2 first to get the complete list of queryable projects (deduplicate by project name).

Function 4: Query Daily Usage for a Specific Feature View

Query daily trends for a specific feature view from StartDate to EndDate (default: last 7 days).

aliyun paifeaturestore list-datasource-feature-views \
  --region <RegionId> \
  --instance-id <InstanceId> \
  --datasource-id <DatasourceId> \
  --start-date <StartDate> \
  --end-date <EndDate> \
  --verbose true \
  --show-storage-usage false \
  --all true \
  --project-name <ProjectName> \
  --name <FeatureViewName> \
  --user-agent AlibabaCloud-Agent-Skills/alibabacloud-pai-feature-store-featuredb-usage-query

Output: Extract the matching feature view from the FeatureViews array.

Error Handling:

• If no matching feature view is found in the response, the feature view does not exist. Recommend querying Function 2 first to get the complete list of queryable feature views.

Step 4: Analyze Results and Calculate Costs

Based on user requirements:

1. Analyze usage trends from the query results
2. Calculate costs Official Pricing (per 10,000 requests):

3. Present insights:

   • Identify high-traffic feature views
   • Highlight usage patterns (spikes, trends)
   • Provide cost breakdown and optimization suggestions if applicable

4. Offer next steps:

   • Ask if the user wants to drill down into specific projects or feature views
   • Offer to calculate costs if not already done
   • Suggest further analysis options

Success Verification

The skill execution is successful if:

1. ✅ Valid PAI-FeatureStore instance is found
2. ✅ FeatureDB datasource is identified or selected
3. ✅ Usage statistics are successfully retrieved
4. ✅ Results are analyzed and presented clearly
5. ✅ Cost calculations are accurate (if requested)
6. ✅ Response language matches user's question language

For detailed verification steps, see references/verification-method.md.

Cleanup

This skill performs read-only operations and does not create any resources. No cleanup is required.

Best Practices

1. Always confirm the RegionId with the user before executing any commands
2. Use pagination properly when listing datasources to ensure all results are retrieved
3. Provide context in analysis: Don't just show numbers—explain trends and patterns
4. Offer actionable insights: Suggest which feature views or projects to investigate further
5. Handle errors gracefully: If a project or feature view doesn't exist, guide the user to list all available options first
6. Calculate costs accurately: Use the correct pricing tier based on the region
7. Keep user engaged: After showing results, ask if they want to drill deeper or analyze specific aspects
8. Decline out-of-scope requests politely: If a request is not about querying usage statistics, politely decline and list supported features

Reference Documentation