All Products
Search
Document Center

Server Load Balancer:CLB monitoring metrics

Last Updated:Jan 15, 2026

You can query monitoring metrics to assess the status, load, and performance of your Classic Load Balancer (CLB) instance. This helps you troubleshoot issues such as request timeouts and throttling. This topic describes the monitoring metrics for CLB.

Usage notes

To call the CloudMonitor API operations, you need to set the Namespace, Period, MetricName, and Dimensions parameters for your CLB instance as follows:

  • Set Namespace to acs_slb_dashboard.

  • Period defaults to 60 (in seconds) and you can also set it to a multiple of 60.

  • Set Dimensions to a JSON string. Example: [{"instanceId":"lb-bp1r92vzpemy099f******", "port":"80", "vip":"121.43.**.**"}]. In Dimensions, set the instanceId field to the ID of your CLB instance, the port field to the port of your CLB instance, and the vip field to its service address.

CLB monitoring metrics

Metrics that can be viewed by using the console, API, or SDK

Layer 4 monitoring metrics

Category

Metric

Description

MetricName

Dimensions

Statistics

Instance connection metrics

New Connections to Instance

The number of new TCP and UDP connections per second on a CLB instance. Unit: Count/s

InstanceNewConnection

instanceId

Average, Minimum, and Maximum

Connections Dropped on Instance

The number of TCP and UDP connections dropped per second on a CLB instance. Unit: Count/s

InstanceDropConnection

instanceId

Average, Minimum, and Maximum

Active Connections to Instance

The number of TCP connections in the ESTABLISHED state on a CLB instance. Unit: Count

Note

A persistent connection can simultaneously transfer multiple file requests.

InstanceActiveConnection

instanceId

Average, Minimum, and Maximum

Inactive Connections to Instance

The number of TCP connections not in the ESTABLISHED state as well as UDP connections on a CLB instance. Unit: Count

InstanceInactiveConnection

instanceId

Average, Minimum, and Maximum

Maximum Concurrent Connections to Instance

The number of TCP and UDP connections processed by a CLB instance. Unit: Count

InstanceMaxConnection

instanceId

Average, Minimum, and Maximum

Instance bandwidth or traffic metrics

Inbound Traffic to Instance

The TCP and UDP traffic received by a CLB instance. Unit: bit/s

InstanceTrafficRX

instanceId

Average, Minimum, and Maximum

Outbound Traffic from Instance

The TCP and UDP traffic sent by a CLB instance per second. Unit: bit/s

InstanceTrafficTX

instanceId

Average, Minimum, and Maximum

Inbound Packets to Instance

The number of TCP and UDP packets received by a CLB instance per second. Unit: Count/s

InstancePacketRX

instanceId

Average, Minimum, and Maximum

Outbound Packets from Instance

The number of TCP and UDP packets sent by a CLB instance per second. Unit: Count/s

InstancePacketTX

instanceId

Average, Minimum, and Maximum

Inbound Traffic Dropped on Instance

The TCP and UDP traffic that a CLB instance fails to receive per second. Unit: bit/s

InstanceDropTrafficRX

instanceId

Average, Minimum, and Maximum

Outbound Traffic Dropped on Instance

The TCP and UDP traffic that a CLB instance fails to send per second. Unit: bit/s

InstanceDropTrafficTX

instanceId

Average, Minimum, and Maximum

Inbound Packets Dropped on Instance

The number of TCP and UDP packets that a CLB instance fails to receive per second. Unit: Count/s

InstanceDropPacketRX

instanceId

Average, Minimum, and Maximum

Outbound Packets Dropped on Instance

The number of TCP and UDP packets that a CLB instance fails to send per second. Unit: Count/s

InstanceDropPacketTX

instanceId

Average, Minimum, and Maximum

Listener connection metrics

New Connections to Listener

The number of new TCP or UDP connections per second on a listener port. Unit: Count/s

NewConnection

instanceId, port, and vip

Average, Minimum, and Maximum

Connections Dropped on Listener

The number of TCP or UDP connections dropped per second on a listener port. Unit: Count/s

DropConnection

instanceId, port, and vip

Average, Minimum, and Maximum

Active Connections to Listener

The number of TCP connections in the ESTABLISHED state on a listener port. Unit: Count

Note

A persistent connection can simultaneously transfer multiple file requests.

ActiveConnection

instanceId, port, and vip

Average, Minimum, and Maximum

Inactive Connections to Listener

The number of TCP connections not in the ESTABLISHED state or UDP connections on a listener port. Unit: Count

InactiveConnection

instanceId, port, and vip

Average, Minimum, and Maximum

Maximum Concurrent Connections to Listener

The number of TCP or UDP connections processed by a listener port. Unit: Count

MaxConnection

instanceId, port, and vip

Average, Minimum, and Maximum

Listener bandwidth or traffic metrics

Inbound Traffic to Listener

The TCP or UDP traffic received by a listener port per second. Unit: bit/s

TrafficRXNew

instanceId, port, and vip

Average, Minimum, and Maximum

Outbound Traffic from Listener

The TCP or UDP traffic sent by a listener port per second. Unit: bit/s

TrafficTXNew

instanceId, port, and vip

Average, Minimum, and Maximum

Inbound Packets to Listener

The number of TCP or UDP packets received by a listener port per second. Unit: Count/s

PacketRX

instanceId, port, and vip

Average, Minimum, and Maximum

Outbound Packets from Listener

The number of TCP or UDP packets sent by a listener port per second. Unit: Count/s

PacketTX

instanceId, port, and vip

Average, Minimum, and Maximum

Inbound Traffic Dropped on Listener

The TCP or UDP traffic that a listener port fails to receive per second. Unit: bit/s

DropTrafficRX

instanceId, port, and vip

Average, Minimum, and Maximum

Outbound Traffic Dropped on Listener

The TCP or UDP traffic that a listener port fails to send per second. Unit: bit/s

DropTrafficTX

instanceId, port, and vip

Average, Minimum, and Maximum

Inbound Packets Dropped on Listener

The number of TCP or UDP packets that a listener port fails to receive per second. Unit: Count/s

DropPacketRX

instanceId, port, and vip

Average, Minimum, and Maximum

Outbound Packets Dropped on Listener

The number of TCP or UDP packets that a listener port fails to send per second. Unit: Count/s

DropPacketTX

instanceId, port, and vip

Average, Minimum, and Maximum

Listener health check metrics

Healthy Backend ECS Instances

The number of healthy ECS instances that serve as backend servers for a listener port.

HeathyServerCount

instanceId, port, and vip

Average, Minimum, and Maximum

Unhealthy ECS Instances

The number of unhealthy ECS instances that serve as backend servers for a listener port.

UnhealthyServerCount

instanceId, port, and vip

Average, Minimum, and Maximum

Layer 7 monitoring metrics

Category

Metric

Description

MetricName

Dimensions

Statistics

Instance connection metrics

Response Time of Layer 7 Instance

The time elapsed from when a client sends a request to when it receives a response through the CLB instance. Unit: ms

InstanceRt

instanceId

Average

UpstreamRT of Layer 7 Instance

The time taken for a CLB instance to forward a request to a backend server and receive a response. Unit: ms

InstanceUpstreamRt

instanceId

Average

Instance request metric

QPS of Layer 7 Instance

The number of HTTP or HTTPS requests that a CLB instance processes per second. Unit: Count/s

InstanceQps

instanceId

Average

Instance status code metrics

2XX Status Codes of Layer 7 Instance

The number of 2xx (Success) status codes returned from a backend server to a client through the CLB instance per second. Unit: Count/s

InstanceStatusCode2xx

instanceId

Average

3XX Status Codes of Layer 7 Instance

The number of 3xx (Redirection) status codes returned from a backend server to a client through the CLB instance per second. Unit: Count/s

InstanceStatusCode3xx

instanceId

Average

4XX Status Codes of Layer 7 Instance

The number of 4xx (Client Error) status codes returned from a backend server to a client through the CLB instance per second. Unit: Count/s

InstanceStatusCode4xx

instanceId

Average

5XX Status Codes of Layer 7 Instance

The number of 5xx (Server Error) status codes returned from a backend server to a client through the CLB instance per second. Unit: Count/s

InstanceStatusCode5xx

instanceId

Average

Other Status Codes of Layer 7 Instance

The number of status codes except 2xx, 3xx, 4xx, and 5xx returned from a backend server to a client through the CLB instance per second. Unit: Count/s

InstanceStatusCodeOther

instanceId

Average

Backend instance status code metrics

4XX Status Codes of Layer 7 Instance

The number of 4xx Client Error status codes returned to a CLB instance from the backend server per second. Unit: Count/s

InstanceUpstreamCode4xx

instanceId

Average

5XX Status Codes of Layer 7 Instance

The number of 5xx Server Error status codes returned to a CLB instance from the backend server per second. Unit: Count/s

InstanceUpstreamCode5xx

instanceId

Average

Listener connection metrics

Response Time of Layer 7 Listener

The time taken for the client to send a request to the backend server and receive a response by using a listener port. Unit: ms

Rt

instanceId, port, and vip

Average

UpstreamRT of Layer 7 Listener

The time taken for a listener port to forward a request to the backend server and receive a response. Unit: ms

UpstreamRt

instanceId, port, and vip

Average

Listener request metric

QPS of Layer 7 Listener

The number of HTTP or HTTPS requests that a listener port processes per second. Unit: Count/s

Qps

instanceId, port, and vip

Average

Listener status code metrics

2XX Status Codes of Layer 7 Listener

The number of 2xx Success status codes returned to the client from the backend server by using a listener port per second. Unit: Count/s

StatusCode2xx

instanceId, port, and vip

Average

3XX Status Codes of Layer 7 Listener

The number of 3xx Redirection status codes returned to the client from the backend server by using a listener port per second. Unit: Count/s

StatusCode3xx

instanceId, port, and vip

Average

4XX Status Codes of Layer 7 Listener

The number of 4xx Client Error status codes returned to the client from the backend server by using a listener port per second. Unit: Count/s

StatusCode4xx

instanceId, port, and vip

Average

5XX Status Codes of Layer 7 Listener

The number of 5xx Server Error status codes returned to the client from the backend server by using a listener port per second. Unit: Count/s

StatusCode5xx

instanceId, port, and vip

Average

Other Status Codes of Layer 7 Listener

The number of status codes except 2xx, 3xx, 4xx, and 5xx returned to the client from the backend server by using a listener port per second. Unit: Count/s

StatusCodeOther

instanceId, port, and vip

Average

Backend listener status code metrics

4XX Status Codes of Layer 7 Listener

The number of 4xx Client Error status codes returned to the CLB instance from the backend server by using a listener port per second. Unit: Count/s

UpstreamCode4xx

instanceId, port, and vip

Average

5XX Status Codes of Layer 7 Listener

The number of 5xx Server Error status codes returned to the CLB instance from the backend server by using a listener port per second. Unit: Count/s

UpstreamCode5xx

instanceId, port, and vip

Average

Metrics that can be viewed by using the API or SDK

Layer 4 monitoring metrics

Category

Metric

Description

MetricName

Dimensions

Statistics

Instance connection metrics

Instance connection utilization

The ratio of established connections to the maximum connections for a CLB instance. Unit: %

InstanceMaxConnectionUtilization

instanceId

Average, Maximum, and Minimum

New instance connection utilization

The ratio of new connections to the maximum allowed new connections for a CLB instance. Unit: %

InstanceNewConnectionUtilization

instanceId

Average, Maximum, and Minimum

Layer 7 monitoring metrics

Category

Metric

Description

MetricName

Dimensions

Statistics

Instance request metric

QPS utilization

The ratio of the current QPS to the maximum QPS of the CLB instance. Unit: %

InstanceQpsUtilization

instanceId

Average, Maximum, and Minimum

FAQ

What is the difference between monitoring data and billing data?

Factor

Monitoring data

Billing data

Calculation method

Data is sampled by the CLB system at a 1-minute granularity. The CloudMonitor service then aggregates these data points and calculates the average value over a 15-minute interval. The metrics displayed in the monitoring console represent this final average.

Billing data is collected at the same granularity. However, the CLB system aggregates these data points into an hourly sum (total value). This hourly total is then sent to the billing system for invoicing. Therefore, billing data reflects the cumulative total over a period, whereas monitoring data reflects an average.

Real-timeliness

Monitoring data is provided in near real-time. However, small delays are inherent in the data collection, processing, and display pipeline. This can lead to minor discrepancies when compared with billing data at any given moment.

The billing system is designed to accommodate processing delays of up to 3 hours. For example, usage data from the 01:00-02:00 hour, which would normally be billed before 03:00, could be delayed and processed as late as 05:00. This intentional latency is a key reason for differences between real-time monitoring and final billing data.

Purpose

To provide operational visibility into the instance's real-time health and performance. It is used to detect anomalies, diagnose issues, and trigger operational responses (e.g., alerts, autoscaling).

To accurately calculate charges based on the actual resources consumed over a billing period. For all financial accounting and invoicing, the billing data is the official source of truth, not the monitoring data.

References