Help Center/ Distributed Cache Service/ User Guide/ Monitoring/ Common Metrics
Updated on 2022-12-12 GMT+08:00
View PDF

Common Metrics

This section describes common Redis metrics.

Table 1 Common metrics

Metric

Description

CPU Usage

This metric indicates the maximum value in each measurement period (minute-level: every minute; second-level: every 5 seconds).

  • For a single-node or master/standby instance, you can view the CPU usage of the instance.
  • For a Proxy Cluster instance, you can view the CPU usage of the Redis Servers and the proxies.
  • For a Redis Cluster instance, you can only view the CPU usage of the Redis Servers.

Memory Usage

This metric measures the memory usage in each measurement period (minute-level: every minute; second-level: every 5 seconds).

  • For a single-node or master/standby instance, you can view the memory usage of the instance.
  • For a Proxy Cluster instance, you can view the memory usage of the instance and the proxies.
  • For a Redis Cluster instance, you can only view the memory usage of the Redis Servers.
NOTICE:

The memory usage does not include the usage of reserved memory.

Connected Clients

This metric indicates the number of instantaneous connected clients, that is, the number of concurrent connections.

This metric does not include the number of connections to the standby nodes of master/standby or cluster instances.

For details about the maximum allowed number of connections, see the "Max. Connections" column of different instance types listed in DCS Instance Specifications.

Ops per Second

This metric indicates the number of operations processed per second.

For details about the maximum allowed number of operations per second, see the "Reference Performance (QPS)" column of different instance types listed in DCS Instance Specifications.

Input Flow

This metric indicates the instantaneous input traffic.

  • The monitoring data on the instance level shows the aggregated input traffic of all nodes.
  • The monitoring data on the node level shows the input traffic of the current node.

Output Flow

This metric indicates the instantaneous output traffic.

  • The monitoring data on the instance level shows the aggregated output traffic of all nodes.
  • The monitoring data on the node level shows the output traffic of the current node.

Bandwidth Usage

This metric indicates the percentage of the used bandwidth to the maximum bandwidth limit.

Bandwidth usage = (Input flow + Output flow)/(2 x Maximum bandwidth) x 100%

Commands Processed

This metric indicates the number of commands processed during the monitoring period. The default monitoring period is 1 minute.

The monitoring period of this metric is different from that of the Ops per Second metric The Ops per Second metric measures the instantaneous number of commands processed. The Commands Processed metric measures the total number of commands processed during the monitoring period.

Flow Control Times

This metric indicates the number of times that the maximum allowed bandwidth is exceeded during the monitoring period.

For details about the maximum allowed bandwidth, see the "Maximum/Assured Bandwidth" column of different instance types listed in DCS Instance Specifications.

Slow Queries

This metric indicates whether slow queries exist on the instance.

For details about the cause of a slow query, see Viewing Redis Slow Queries.
Parent topic: Monitoring