Fine-grained monitoring

Updated at:
Copy as MD

Classic Load Balancer (CLB) provides fine-grained monitoring using Simple Log Service. You can use metrics from a layer 4 listener to view per-second data for your CLB instance, including traffic, new connections, and maximum connections. This enables more detailed service monitoring and more effective troubleshooting.

Supported regions

For a list of regions where fine-grained monitoring is available, see the following tables.

Area

Region

China

China (Hangzhou), China (Shanghai), China (Qingdao), China (Beijing), China (Shenzhen), China (Hohhot), China (Heyuan), China (Guangzhou), China (Chengdu), Hong Kong (China)

Asia Pacific

Japan (Tokyo), Malaysia (Kuala Lumpur), Singapore, Indonesia (Jakarta)

Europe and Americas

Germany (Frankfurt), UK (London), US (Virginia), US (Silicon Valley)

Middle East

UAE (Dubai)

Billing

After a CLB instance sends logs to Simple Log Service, Simple Log Service charges you for storage space, read traffic, requests, data transformation, and data shipping. For more information, see Simple Log Service Pricing.

Prerequisites

Enable fine-grained monitoring

  1. Log on to the CLB console.

  2. In the top navigation bar, select the region where the CLB instance is located.

  3. On the Instances page, click the ID of the target instance, and then click the Fine-grained Monitoring tab.

  4. Enable fine-grained monitoring for the current region.

    If fine-grained monitoring is already enabled in the current region, skip this step.

    1. On the Fine-grained Monitoring tab, click Enable Fine-grained Monitoring.

    2. In the Enable Fine-grained Monitoring for SLB dialog box, select or enter a Project and a MetricStore, select I understand the above information., and then click OK.

      Important
      • This operation automatically creates a RAM role named AliyunServiceRoleForSlbLogDelivery. This role authorizes Server Load Balancer to push per-second monitoring metrics to Simple Log Service.

      • Do not revoke the authorization or delete the RAM role. Otherwise, Server Load Balancer cannot send per-second monitoring metrics to Simple Log Service.

  5. Enable fine-grained monitoring for the target TCP listener or UDP listener.

    1. Open the Fine-grained Monitoring > Settings tab.

    2. On the Settings tab, turn on the Fine-grained Monitoring switch for the target TCP listener or UDP listener.

    Note

    After you enable the fine-grained monitoring feature, Simple Log Service receives monitoring data from CLB with a delay of about 5 seconds.

Next steps

After Simple Log Service collects per-second monitoring data from a CLB instance, you can query, analyze, download, ship, and process the data, or configure alerts in the Simple Log Service console. For more information, see Common operations for cloud product logs.

Metrics

The metrics in this topic use the time series data format. For more information, see Time series data (metric).

You can use PromQL or SQL statements to query and analyze data. For more information, see Introduction to time series data query and analysis.

Table 1. Metrics

Metric

Description

actConnsPS

Active connections per second.

connsPS

New connections per second.

dropConnPS

Dropped connections per second.

failConnPS

Failed connections per second.

inActConnPS

Inactive connections per second.

inBitsPS

Inbound bits per second. Unit: bit/s.

inDropBitsPS

Dropped inbound bits per second. Unit: bit/s.

inDropPktsPS

Dropped inbound packets per second.

inPktsPS

Inbound packets per second.

aclDropBitsPS

Inbound bits dropped by access control lists (ACLs) per second. Unit: bit/s.

aclDropPktsPS

Inbound packets dropped by ACLs per second.

maxConnsPs

Concurrent connections per second.

outBitsPS

Outbound bits per second. Unit: bit/s.

outDropBitsPS

Dropped outbound bits per second. Unit: bit/s.

outDropPktsPS

Dropped outbound packets per second.

Table 2. Labels

Label

Description

lbId

The ID of the SLB instance.

listenerId

The ID of the listener.

protocol

Protocol

vip

The IP address used by the listener.

vport

The port used by the listener.

Related documents

Layer 4 per-second monitoring metrics for Server Load Balancer