How to set spending limits and alerts in Gate.AI

This tutorial introduces how to configure consumption guardrails and alert notifications using Gate.AI to achieve organization-level AI cost control. After completing the configuration, you will be able to restrict the model consumption quotas for your organization or members, and automatically receive alert notifications when preset thresholds are reached or abnormal usage occurs.

After completing this tutorial, you will be able to:

• Configure organization-level consumption limits
• Configure member-level consumption limits
• Set quota reset cycles
• Enable organization alert push notifications
• Configure Webhook callback addresses
• Receive alerts for quota thresholds and abnormal usage
• Establish basic AI cost governance mechanisms

Before starting, ensure the following conditions are met:

• You have a Gate.AI enterprise account
• You have organization administrator permissions
• You have created an organization
• You have created at least one member account or API key
• You have a server endpoint capable of receiving HTTPS Webhook requests

Step 1: Access the Guardrails Configuration Page

Consumption quota control is implemented through Gate.AI's "Guardrails" feature.

Navigation path:

Plain Console → Settings → Guardrails

After entering the page, you can view the current quota rules created for the organization, including:

• Organization limits
• Member limits
• API key restrictions
• RPM limits
• Budget cycle

If no rules have been created yet, click the "Add Guardrail Configuration" button in the top right to start configuring.

Step 2: Set Organization Consumption Limit

Organization limits are used to control the maximum spending amount for the entire organization within a specified cycle.

Click:

Plain Add Guardrail Configuration → Organization Limit

Example configuration:

Plain Name: Organization Guardrail

Budget: 30 USD

Reset cycle: Daily

After saving, the system will continuously track the total consumption amount of the organization within each statistical cycle.

Expected result:

Plain Organization limit: 30 USD / Daily

Organization limits are suitable as the first layer of enterprise-level cost control. For teams just starting to use AI services, it is recommended to set a lower limit and adjust it gradually based on actual usage.

If you are building a comprehensive enterprise cost governance system, it is advisable to first understand the basic principles of AI Cost Management before formulating budget strategies.

Step 3: Set Member Consumption Limit

Organization limits control overall costs, while member limits can further restrict resource consumption by individual users.

Click:

Plain Add Guardrail Configuration → Member Limit

Example configuration:

Plain Name: Team A Guardrail

Budget: 20 USD

Reset cycle: Daily

After saving, the system will track consumption amounts separately for members who meet the rules.

Expected result:

Plain Member limit: 20 USD / Daily

For collaborative teams, it is recommended to configure both organization limits and member limits simultaneously to prevent individual users from consuming large resources due to misoperations or abnormal requests.

Step 4: Configure Alert Push Address

Consumption limit controls expenses, while alert notifications are used to proactively detect abnormal situations.

Navigation path:

Plain Console → Settings → Push Notifications

In the "Callback URL" area, fill in a Webhook address capable of receiving HTTPS requests:

Example:

Plain

Click Save, and Gate.AI will send a verification request to this address.

Your server needs to return:

JSON { "status": "ok" }

Expected result:

Plain Callback address verified Webhook verification successful

It is recommended to use an internal enterprise alert service as a unified receiving endpoint, then forward to Feishu, WeChat Work, Slack, or email systems.

Step 5: Enable Alert Push Notifications

After configuring the Webhook, you can enable the alert push feature.

Navigation path:

Plain Console → Settings → Push Notifications → Alert Push

Turn on the "Alert Push" switch.

Expected result:

Plain Alert push: Enabled

Once enabled, when the organization’s quota reaches the preset threshold or abnormal usage occurs, the system will automatically send a JSON message to the configured Webhook address.

Typical scenarios include:

• Organization quota nearing the limit
• Abnormal growth in member consumption
• Abnormal increase in API request volume
• High-frequency calls triggering limit policies

Step 6: Verify if Alerts Are Triggered Properly

After completing the configuration, it is recommended to perform a verification test.

Verification steps:

1. Create a test consumption scenario.
2. Trigger quota increase.
3. Check if consumption statistics are updated.
4. Review Webhook service logs.
5. Confirm receipt of Gate.AI push event.

Example event:

JSON { "event": "org_quota_threshold", "org_id": 12345, "rule_name": "Quota System Warning" }

Expected result:

Plain Webhook received successfully

If you can receive the JSON message normally, it indicates that the alert mechanism is working.

What is the difference between organization limits and member limits?

Gate.AI provides multi-level quota control mechanisms.

It is recommended that enterprises enable both organization limits and member limits to form a dual-layer cost control system.

How to recommend configuring consumption guardrails?

Different team sizes require different quota strategies.

Actual quotas should be adjusted based on model types, call frequency, and business scenarios.

If your organization needs to manage multiple model providers simultaneously, it is recommended to integrate AI Gateway architecture for unified management of costs, permissions, and model access policies.

What to do if alert notifications are not received?

If the system has reached alert conditions but notifications are not received, check the following:

Check Webhook address

Ensure:

• Using HTTPS
• Domain is accessible
• No firewall blocking

Check verification request

Ensure the server returns:

JSON { "status": "ok" }

Check alert push status

Ensure:

Plain Alert push: Enabled

and not:

Plain Alert push: Disabled

Check server logs

Confirm whether POST requests from Gate.AI have been received.

What can be done next?

After completing the consumption guardrail configuration, you can further enhance your organization’s AI governance capabilities.

As your team grows, you can further configure RBAC permission management to isolate permissions among members, teams, and applications.

For large organizations, it is also possible to establish organization-level AI governance strategies to centrally manage budgets, audit logs, model access permissions, and security policies.

Additionally, you can explore:

• API key lifecycle management
• Multi-model unified routing
• Team cost attribution analysis
• Enterprise audit log systems

FAQ

Q: Should organization limits and member limits be set simultaneously?

A: It is recommended to set both. Organization limits control overall costs, while member limits restrict individual resource consumption. Combining both provides a more comprehensive cost protection mechanism.

Q: Why does consumption still occur after reaching the quota?

A: Please check whether the guardrail rules have been successfully saved, and whether the corresponding members or API keys are correctly included in the restrictions.

Q: Why haven’t you received alert notifications?

A: Please verify whether the Webhook address is accessible, whether alert push is enabled, and whether the server correctly returns verification results.

Q: Must Webhook use HTTPS?

A: Yes. The current callback address only supports HTTPS protocol to ensure the security of notification transmission.

Q: Can alert notifications be sent directly to Feishu or WeChat Work?

A: Gate.AI currently pushes JSON data via Webhook. You can receive the message on your server and then forward it to Feishu, WeChat Work, Slack, or email systems.