How is SLB billed?

See Billing method.

Is the inbound traffic of SLB billed?

No. Only the outbound traffic of SLB is billed. For more information, see Network traffic flow.

Is traffic generated by health checks billed?

No. The traffic generated by health checks of SLB is not included in the billed traffic.

Is the billing method of an ECS instance changed after it is added to the backend server pool of SLB?

No. The billing method of an ECS instance does not change whether or not you associate it with an SLB instance. SLB and ECS instances are billed separately based on usage.

Is traffic generated by unexpected traffic spikes, or malicious activity such as a DDoS attack, billed?

Yes. However, SLB can work with Alibaba Cloud Security to protect your SLB instances and minimize the billed traffic caused by malicious attacks. A latency of several seconds may happen from the time when the attack traffic reaches the scrubbing or blackholing threshold to the time when Alibaba Cloud Security starts scrubbing or blackholing. During this period, responses may be sent to attacks and this incurs fees. Also, such attacks consume the bandwidth resources of SLB instances.

If all the backend ECS instances of an SLB instance are stopped or removed, does the SLB instance continue to be billed?

Yes. Fees are charged for the following reasons:

  • Traffic-based billing

    In the case of traffic-based billing, no traffic fee is incurred when an SLB instance is stopped, released, or not accessed.

    SLB is a traffic distribution and control service in front of backend ECS instances. SLB provides services through its service address. If all backend ECS instances are stopped, but the SLB instance is not stopped, inbound traffic can still reach the service address of the SLB instance. In this case, the SLB instance responds when it discovers that no backend ECS instances are available after performing health checks.

    For layer-4 SLB, only three-way handshake packets are returned. For layer-7 SLB, a Tengine 503 error page is prompted because the service is provided by Tengine. If requests keeps reaching SLB, SLB will respond continuously. The response traffic is billed.

    This also applies to SLB instances with no ECS instances added. Therefore, we recommend that you stop the SLB instance if you no longer need it.

Is a specification fee charged to intranet SLB instances?

  • If the intranet SLB instance is a shared-performance instance, no specification fee is charged.
  • If the intranet SLB instance is a guaranteed-performance instance, a specification fee is charged.

    The calculation method of specification fees for intranet SLB instances is the same as that for Internet SLB instances. No instance fee or traffic fee is charged for intranet instances.

Are the traffic fees and instance fees charged for guaranteed-performance instances the same as those charged for shared-performance instances?


Is an extra fee included for shared-performance instances after Alibaba Cloud starts charging for the specification fee?


Extra fees are not charged for shared-performance instances unless you change them to guaranteed-performance instances.

Why is the monitoring data of SLB instances different from the data recorded on my bills?

  • Take traffic data as an example. The monitoring data displayed in the SLB console are average values. SLB collects monitoring data at one minute intervals and reports the data to CloudMonitor, which then calculates the average value at every 15-minute intervals. However, the data reported on your bills are accumulated values. Specifically, SLB collects traffic data at one minute intervals, and reports the accumulated values at every one hour intervals to the billing system.

    This means that the data reported to your billing system are accumulated values calculated at one hour intervals, whereas the monitoring data displayed in the console is of average values calculated at every 15 minute interval. The different intervals used in the two calculations result in a discrepancy in monitoring data and billing data.

  • Monitoring data and billing data have different latency requirements. SLB provides real-time monitoring data. However, a short delay may occur during data collection, calculation, and display. Such a delay can cause a discrepancy between monitoring and billing data. Billing data can be recorded after a maximum delay of three hours. For example, billing data generated between 01:00 and 02:00 is normally reported to the billing system before 03:00, but due to a delay, may be reported to the billing system at 05:00, resulting in a discrepancy between billing data and monitoring data.
  • Monitoring data and billing data have different purposes. Specifically, the purpose of monitoring is to help you observe if instances are running normally. If the instances are running abnormally, you can take measures to resolve problems in a timely manner. The purpose of billing data is to generate bills based on the actual usage of resources under your account.

Why is the value of the traffic actually used by HTTPS protocol greater than the value recorded on my bills?

Some traffic is generated due to handshakes in HTTPS transactions. Therefore, the actual traffic value is higher than the value recorded on your bills.