Server Load Balancer:CLB billing FAQ

What are the billable items for a CLB instance?

A pay-as-you-go CLB instance may incur the following fees. The applicable items depend on the network type and billing method of the instance.

Starting from December 1, 2024, 00:00:00 (UTC+8), "instance fee" was renamed "public IP retention fee" and a new "instance fee" billable item was added. CLB instances created before that date are exempt from instance fees until November 30, 2026, 23:59:59 (UTC+8). For details, see CLB billing adjustments and Extension of the instance fee waiver period for CLB.

How are guaranteed-performance instances billed?

Pay-by-specification (Internet-facing):

Total fee = Public IP address retention fee + Instance fee + Data transfer or bandwidth fee + Specification fee

The specification fee is fixed based on the selected specification. For example, if you select the super large I (slb.s3.large) specification, you are charged for that specification even if actual usage reaches only the level of a high-performance I (slb.s3.small) instance.

Pay-by-LCU (Internet-facing):

Total fee = Public IP address retention fee + Instance fee + Data transfer fee + LCU fee

Performance scales automatically with usage. No specification selection is required. The LCU fee is charged hourly based on actual consumption.

For full billing details, see Pay-as-you-go billing for CLB.

Do shared-resource instances incur specification fees?

No. Specification fees apply only to guaranteed-performance instances.

The billable items vary by network type, metering method, and instance type:

CLB instances created before December 1, 2024, 00:00:00 (UTC+8) are exempt from instance fees until November 30, 2026, 23:59:59 (UTC+8). See Extension of the instance fee waiver period for CLB.

Are internal CLB instances charged specification fees?

It depends on the instance type:

• Shared-resource instances (discontinued): No specification fee.

• Guaranteed-performance instances: A specification fee applies, using the same billing rules as Internet-facing instances.

For details, see Pay-as-you-go billing for CLB.

How can I estimate LCU usage and fees?

Two options are available:

• Manual estimation: Use your business metrics and the examples in Pay-as-you-go billing for CLB.

• Calculator: Use the CLB LCU calculator to estimate consumption automatically.

Does adding Elastic Compute Service (ECS) instances to a CLB backend affect CLB billing?

No. CLB and ECS are billed and settled separately. Adding ECS instances to the server group does not change the billing rules for the CLB instance, regardless of how those ECS instances are billed.

Are fees still charged after I stop a CLB instance?

Yes. A CLB instance stops accruing charges only after it is released.

Fees that continue after the instance is stopped:

• Public IP address retention fee: Charged as long as the public IP address exists, even if not in use.

• Instance fee: Charged as long as the instance exists and has not been released.

• Specification fee: Resources remain reserved so the instance can be restarted at any time. The specification fee continues until the instance is released.

• Bandwidth fee (pay-by-bandwidth only): Charged based on the purchased bandwidth until the instance is released. Not affected by instance status or traffic.

Fees that stop when the instance is stopped:

• Data transfer fee (pay-by-data-transfer only): No traffic is forwarded when the instance is stopped, so no data transfer fee accrues.

• LCU fee: No LCUs are consumed when the instance is stopped.

Why did I receive a bill after releasing my CLB instance?

Two reasons are common:

• Billing delay: Pay-as-you-go bills are generated with an approximate 3-hour delay. For example, if you release an instance at 14:30, the bill covering 14:00–15:00 may not appear until around 18:00. This is a normal billing cycle delay, not a duplicate charge.

• Associated elastic IP address (EIP) not released: If an internal CLB instance was bound to a separately purchased EIP, releasing the CLB instance does not automatically release the EIP. The EIP continues to incur fees until it is released.

How long does it take for service to resume after topping up an overdue account?

For a subscription instance, the service resumes immediately after renewal.

For a pay-as-you-go instance that is suspended due to an overdue payment, service resumes immediately after the overdue payment is cleared. If service does not resume promptly, verify that the payment was processed successfully, or submit a ticket to contact technical support.

Is inbound traffic to CLB billed?

No. CLB charges only for outbound traffic. Inbound traffic is free. For details on CLB traffic paths, see Inbound traffic path and Outbound traffic path.

Do internal CLB instances incur data transfer fees?

No. The billable items for an internal CLB instance include only the instance fee and the LCU fee (or specification fee). Public IP address retention fees, data transfer fees, and bandwidth fees do not apply.

If you bind a separately purchased EIP to the CLB instance, that EIP incurs public network data transfer fees. These charges appear on the EIP bill, not the CLB bill.

Is health check traffic billed?

No. Traffic generated by CLB health checks is not included in data transfer fees.

Is attack traffic billed?

Partially. Alibaba Cloud Security provides traffic scrubbing and blackholing to protect against attacks. A short delay occurs between when an attack reaches the scrubbing or blackholing threshold and when protection activates. Traffic responses sent during this delay are billed and consume CLB bandwidth.

Why is the actual HTTPS traffic volume higher than what appears on my bill?

The HTTPS protocol uses additional traffic for TLS handshakes. This handshake overhead is not included in billed traffic, so the actual traffic volume is higher than the billed amount.

Are public network fees still charged when all backend ECS instances are stopped or removed?

Public IP address retention fees, instance fees, and specification fees continue regardless of backend ECS instance status.

For traffic fees:

• Pay-by-data-transfer: Traffic fees depend on whether the CLB instance continues to receive requests. Even with all backend ECS instances stopped, the CLB endpoint remains active. Incoming requests trigger health check responses — three-way handshake packets for Layer 4, or a Tengine 503 error page for Layer 7. This response traffic is billed. To avoid unnecessary data transfer fees, stop or release CLB instances that are no longer in use.

• Pay-by-bandwidth: Fees are fixed based on purchased bandwidth and continue until the instance is released, regardless of instance status or traffic volume.

Which CLB traffic can be offset by Data Transfer Plans?

• IPv4 traffic from the public IP address of a Internet-facing CLB instance using pay-by-data-transfer metering.

• IPv4 traffic (excluding BGP (Multi-ISP) Pro EIP traffic) from a pay-by-data-transfer EIP bound to an internal CLB instance in a Virtual Private Cloud (VPC).

For more information, see Data Transfer Plan.

Does changing from pay-by-specification to pay-by-LCU affect my service?

It depends on the current specification:

• Super large I (slb.s3.large) or lower: The change does not affect your service.

• Higher than super large I: Switching may degrade instance performance. Proceed with caution.

  Note

  • By default, CLB specifications higher than super large I are not available. For Layer 4 load balancing, if you require more concurrent connections, select a Network Load Balancer (NLB) product. For Layer 7 load balancing, if you require a higher number of queries per second (QPS), select an Application Load Balancer (ALB) product.

    • For more information about NLB, see What is NLB?

    • For more information about ALB, see What is ALB?

  • Starting June 1, 2025, 00:00:00 (UTC+8), pay-by-specification CLB instances are no longer available for purchase, and pay-by-LCU instances cannot be converted to pay-by-specification instances any more. For more details, see End of sale for pay-by-specification CLB instances.

Pay-by-LCU instances also have performance limits. For details, see Switch from pay-by-specification to pay-by-usage.

Does upgrading or downgrading an instance specification affect my service?

No. An upgrade or downgrade does not interrupt your service. However, if the change involves a network type or traffic configuration change, the new configuration takes effect at 00:00:00 the next day. No additional upgrade or downgrade operations can be performed until the current change is complete.

How do I view the fees for a specific CLB instance?

1. Log on to the Expenses and Costs console.

2. In the Resource Instance Name/ID field, enter the instance name or ID. Obtain the ID from the Server Load Balancer console.

3. Click Search.

The console also supports Customize Columns and Export features via the icons in the upper-right corner of the table.

How do I view CLB traffic usage?

Two methods are available:

• Real-time monitoring: Log on to the CloudMonitor console, locate the target instance, and view traffic monitoring data.

• Billed traffic from your bill: Log on to the Expenses and Costs console, filter for the target instance, and view usage details.

Discrepancies between monitoring data and bill data are normal. They are caused by different collection methods and reporting schedules. Your bill is calculated based on bill data, not monitoring data.

Why does monitoring data differ from my bill?

Three factors cause this discrepancy:

• Different collection pipelines: CLB console traffic data is collected at 1-minute intervals and reported to CloudMonitor. Bill data is also collected at 1-minute intervals, then aggregated into an hourly accumulated value and reported to the billing system. These two pipelines are independent and not directly comparable.

• Reporting delay tolerance: The billing system allows bill data to be reported up to 3 hours late. For example, usage data from 01:00–02:00 must be reported to the billing system by 05:00. This flexible window can cause bill data to appear later than monitoring data.

• Different purposes: Monitoring data is designed to detect operational anomalies in near real time. Bill data is designed for accurate charge calculation based on actual resource consumption. For billing reconciliation, always use bill data as the authoritative source.

How do I investigate a sudden spike in data transfer fees?

1. Identify the spike: Log on to the CloudMonitor console and view traffic monitoring data for the CLB instance to pinpoint when the spike occurred.

2. Find the traffic source:

   • Review access logs on the backend servers to identify abnormal requests. CLB supports configuring backend servers to obtain real client IP addresses.

   • Analyze request sources using CLB access logs. Enable access logging in advance if it is not already configured.

3. Determine the cause: Check for malicious attacks, web crawler activity, or sudden increases in legitimate business traffic.

4. Protect against recurrence:

   • Use Web Application Firewall (WAF) to block malicious traffic.

   • Apply access control lists (ACLs) to restrict source IP addresses.

   • Configure bandwidth monitoring alerts to detect future anomalies promptly.