Help Center/ Distributed Cache Service/ User Guide/ Viewing Monitoring Metrics and Configuring Alarms/ DCS Metrics
Updated on 2025-11-03 GMT+08:00
View PDF

DCS Metrics

Introduction

This section describes DCS metrics reported to Cloud Eye as well as their namespaces and dimensions. You can use the Cloud Eye console or call APIs to query the DCS metrics and alarms.

Cloud Eye supports up to 4 hierarchical dimensions, numbered from 0, with 3 as the deepest. For example, if the metric dimension is dcs_instance_id,dcs_cluster_redis_node, then dcs_instance_id is level 0 and dcs_cluster_redis_node is level 1.

Table 1 compares instance metrics.

Table 1 Monitoring dimensions for different instance types

Instance Type

Instance Monitoring

Redis Server Monitoring

Proxy Monitoring

Single-node

Supported

The monitoring on the instance dimension is conducted on the Redis Server.

N/A

N/A

Master/standby

Supported

The master node is monitored.

Supported

The master and standby nodes are monitored.

N/A

Read/write splitting

Supported

The master node is monitored.

Supported

The master and standby nodes are monitored.

Supported

Each proxy is monitored.

Proxy Cluster

Supported

The monitoring data is the aggregated master node data.

Supported

Each shard is monitored.

Supported

Each proxy is monitored.

Redis Cluster

Supported

The monitoring data is the aggregated master node data.

Supported

Each shard is monitored.

N/A

Namespace

SYS.DCS

DCS Redis 4.0 and Later Instance Metrics

  • Dimensions lists the metric dimensions.
  • The monitoring data is the aggregated master node data.
  • Some metrics are aggregated from the master and replica nodes. For details, see "Metric Description" in Table 2.
Table 2 DCS Redis 4.0 and later instance metrics

Metric ID

Metric Name

Metric Description

Value Range

Unit

Conversion Rule

Dimension

Monitoring Period (Raw Data)

cpu_usage

Maximum CPU usage

The monitored object's maximum CPU usage among multiple sampling values in a monitoring period

Supported for single-node, master/standby, and read/write splitting instances.

0–100

%

N/A

dcs_instance_id

1 minute

cpu_avg_usage

Average CPU Usage

The monitored object's average CPU usage of multiple sampling values in a monitoring period

Supported for single-node, master/standby, and read/write splitting instances.

0–100

%

N/A

dcs_instance_id

1 minute

command_max_delay

Maximum Command Latency

Maximum latency of commands

≥ 0

ms

N/A

dcs_instance_id

1 minute

total_connections_received

New Connections

Number of connections received during the monitoring period. Includes connections from replicas and established for system monitoring, configuration synchronization, and services

≥ 0

N/A

N/A

dcs_instance_id

1 minute

is_slow_log_exist

Slow Query Logs

Existence of slow query logs in the instance

Slow queries caused by the MIGRATE, SLAVEOF, CONFIG, BGSAVE, and BGREWRITEAOF commands are not counted.
  • 1: yes
  • 0: no

N/A

N/A

dcs_instance_id

1 minute

memory_usage

Memory Usage

Memory consumed by the monitored object

0–100

%

N/A

dcs_instance_id

1 minute

expires

Keys With an Expiration

Number of keys with an expiration in Redis

≥ 0

N/A

N/A

dcs_instance_id

1 minute

keyspace_hits_perc

Hit Rate

Ratio of the number of Redis cache hits to the number of lookups. Calculation: keyspace_hits/(keyspace_hits + keyspace_misses)

Aggregated from the master and replica nodes. If no read command is performed within a monitoring period, the ratio is 0.

0–100

%

N/A

dcs_instance_id

1 minute

used_memory

Used Memory

Total number of bytes used by the Redis server

≥ 0

byte

1024(IEC)

dcs_instance_id

1 minute

used_memory_dataset

Used Memory Dataset

Dataset memory that the Redis server has used

≥ 0

byte

1024(IEC)

dcs_instance_id

1 minute

used_memory_dataset_perc

Used Memory Dataset Ratio

Percentage of dataset memory that server has used

Aggregated from the master and replica nodes.

0–100

%

N/A

dcs_instance_id

1 minute

used_memory_rss

Used Memory RSS

Resident set size (RSS) memory that the Redis server has used, which is the memory that actually resides in the memory, including all stack and heap memory but not swapped-out memory

≥ 0

byte

1024(IEC)

dcs_instance_id

1 minute

instantaneous_ops

Ops per Second

Number of commands processed per second

≥ 0

N/A

N/A

dcs_instance_id

1 minute

keyspace_misses

Keyspace Misses

Number of failed lookups of keys in the main dictionary during the monitoring period

Aggregated from the master and replica nodes.

≥ 0

N/A

N/A

dcs_instance_id

1 minute

keys

Keys

Number of keys in Redis

≥ 0

N/A

N/A

dcs_instance_id

1 minute

blocked_clients

Blocked Clients

Number of clients suspended by block operations

≥ 0

N/A

N/A

dcs_instance_id

1 minute

connected_clients

Connected Clients

Number of connected clients. Includes connections established for system monitoring, configuration synchronization, and services.

≥ 0

N/A

N/A

dcs_instance_id

1 minute

del

DEL

Number of DEL commands processed per second

0–500,000

Count/s

N/A

dcs_instance_id

1 minute

evicted_keys

Evicted Keys

Number of keys that have been evicted and deleted during the monitoring period

Aggregated from the master and replica nodes.

≥ 0

N/A

N/A

dcs_instance_id

1 minute

expire

EXPIRE

Number of EXPIRE commands processed per second

0–500,000

Count/s

N/A

dcs_instance_id

1 minute

expired_keys

Expired Keys

Number of keys that have expired and been deleted during the monitoring period

Aggregated from the master and replica nodes.

≥ 0

N/A

N/A

dcs_instance_id

1 minute

get

GET

Number of GET commands processed per second

Aggregated from the master and replica nodes.

0–500,000

Count/s

N/A

dcs_instance_id

1 minute

hdel

HDEL

Number of HDEL commands processed per second

0–500,000

Count/s

N/A

dcs_instance_id

1 minute

hget

HGET

Number of HGET commands processed per second

Aggregated from the master and replica nodes.

0–500,000

Count/s

N/A

dcs_instance_id

1 minute

hmget

HMGET

Number of HMGET commands processed per second

Aggregated from the master and replica nodes.

0–500,000

Count/s

N/A

dcs_instance_id

1 minute

hmset

HMSET

Number of HMSET commands processed per second

0–500,000

Count/s

N/A

dcs_instance_id

1 minute

hset

HSET

Number of HSET commands processed per second

0–500,000

Count/s

N/A

dcs_instance_id

1 minute

instantaneous_input_kbps

Input Flow

Instantaneous input traffic

≥ 0

KiB/s

1024(IEC)

dcs_instance_id

1 minute

instantaneous_output_kbps

Output Flow

Instantaneous output traffic

≥ 0

KiB/s

1024(IEC)

dcs_instance_id

1 minute

memory_frag_ratio

Memory Fragmentation Ratio

Current memory fragmentation

≥ 0

N/A

N/A

dcs_instance_id

1 minute

mget

MGET

Number of MGET commands processed per second

Aggregated from the master and replica nodes.

0–500,000

Count/s

N/A

dcs_instance_id

1 minute

mset

MSET

Number of MSET commands processed per second

0–500,000

Count/s

N/A

dcs_instance_id

1 minute

pubsub_channels

PubSub Channels

Number of Pub/Sub channels

≥ 0

N/A

N/A

dcs_instance_id

1 minute

pubsub_patterns

PubSub Patterns

Number of Pub/Sub patterns

≥ 0

N/A

N/A

dcs_instance_id

1 minute

set

SET

Number of SET commands processed per second

0–500,000

Count/s

N/A

dcs_instance_id

1 minute

used_memory_lua

Used Memory Lua

Number of bytes used by the Lua engine

≥ 0

byte

1024(IEC)

dcs_instance_id

1 minute

used_memory_peak

Used Memory Peak

Peak memory consumed by Redis since the Redis server last started

≥ 0

byte

1024(IEC)

dcs_instance_id

1 minute

sadd

Sadd

Number of SADD commands processed per second

0–500,000

Count/s

N/A

dcs_instance_id

1 minute

smembers

Smembers

Number of SMEMBERS commands processed per second

Aggregated from the master and replica nodes.

0–500,000

Count/s

N/A

dcs_instance_id

1 minute

scan

SCAN

Number of SCAN operations per second

0–500,000

Count/s

N/A

dcs_instance_id

1 minute

setex

SETEX

Number of SETEX operations per second

0–500,000

Count/s

N/A

dcs_instance_id

1 minute

rx_controlled

Flow Control Times

Number of times that client requests are controlled in a period. This metric is incremented by 1 each time a client request is controlled.

If the value is greater than 0, the consumed bandwidth exceeds the upper limit and flow control is triggered on a node. The node suspends client commands temporarily.

≥ 0

Count

N/A

dcs_instance_id

1 minute

bandwidth_usage

Bandwidth Usage

Percentage of the used bandwidth to the maximum bandwidth limit

0–200

%

N/A

dcs_instance_id

1 minute

command_max_rt

Maximum Latency

Maximum delay from when the node receives commands to when it responds

Supported for single-node instances.

≥ 0

μs

N/A

dcs_instance_id

1 minute

command_avg_rt

Average Latency

Average delay from when the node receives commands to when it responds

Supported for single-node instances.

≥ 0

μs

N/A

dcs_instance_id

1 minute

Redis Server Metrics of DCS Redis Instances

  • Data nodes of master/standby, read/write splitting, and cluster DCS Redis instances can be monitored. Monitoring metrics are available for each node.
  • Dimensions lists the metric dimensions.
Table 3 Redis Server metrics

Metric ID

Metric Name

Metric Description

Value Range

Unit

Conversion Rule

Dimension

Monitoring Period (Raw Data)

cpu_usage

Maximum CPU usage

The monitored object's maximum CPU usage among multiple sampling values in a monitoring period

0–100

%

N/A

dcs_cluster_redis_node

1 minute

cpu_avg_usage

Average CPU Usage

The monitored object's average CPU usage of multiple sampling values in a monitoring period

0–100

%

N/A

dcs_cluster_redis_node

1 minute

memory_usage

Memory Usage

Memory consumed by the monitored object

0–100

%

N/A

dcs_cluster_redis_node

1 minute

connected_clients

Connected Clients

Number of connected clients. Includes connections established for system monitoring, configuration synchronization, and services.

≥ 0

N/A

N/A

dcs_cluster_redis_node

1 minute

client_longest_out_list

Client Longest Output List

Longest output list among current client connections

Only supported by master/standby, read/write splitting, and cluster Redis 4.0 instances.

≥ 0

N/A

N/A

dcs_cluster_redis_node

1 minute

client_biggest_in_buf

Client Biggest Input Buf

Maximum input data length among current client connections

Only supported by master/standby, read/write splitting, and cluster Redis 4.0 instances.

≥ 0

byte

1024(IEC)

dcs_cluster_redis_node

1 minute

blocked_clients

Blocked Clients

Number of clients suspended by block operations such as BLPOP, BRPOP, and BRPOPLPUSH

≥ 0

N/A

N/A

dcs_cluster_redis_node

1 minute

used_memory

Used Memory

Total number of bytes used by the Redis server

≥ 0

byte

1024(IEC)

dcs_cluster_redis_node

1 minute

used_memory_rss

Used Memory RSS

RSS memory that the Redis server has used, which includes all stack and heap memory but not swapped-out memory

≥ 0

byte

1024(IEC)

dcs_cluster_redis_node

1 minute

used_memory_peak

Used Memory Peak

Peak memory consumed by Redis since the Redis server last started

≥ 0

byte

1024(IEC)

dcs_cluster_redis_node

1 minute

used_memory_lua

Used Memory Lua

Number of bytes used by the Lua engine

≥ 0

byte

1024(IEC)

dcs_cluster_redis_node

1 minute

memory_frag_ratio

Memory Fragmentation Ratio

Current memory fragmentation, which is the ratio between used_memory_rss/used_memory.

≥ 0

N/A

N/A

dcs_cluster_redis_node

1 minute

total_connections_received

New Connections

Number of connections received during the monitoring period. Includes connections from replicas and established for system monitoring, configuration synchronization, and services

≥ 0

N/A

N/A

dcs_cluster_redis_node

1 minute

total_commands_processed

Commands Processed

Number of commands processed during the monitoring period

≥ 0

N/A

N/A

dcs_cluster_redis_node

1 minute

instantaneous_ops

Ops per Second

Number of commands processed per second

≥ 0

N/A

N/A

dcs_cluster_redis_node

1 minute

total_net_input_bytes

Network Input Bytes

Number of bytes received during the monitoring period

≥ 0

byte

1024(IEC)

dcs_cluster_redis_node

1 minute

total_net_output_bytes

Network Output Bytes

Number of bytes sent during the monitoring period

≥ 0

byte

1024(IEC)

dcs_cluster_redis_node

1 minute

instantaneous_input_kbps

Input Flow

Instantaneous input traffic

≥ 0

KiB/s

1024(IEC)

dcs_cluster_redis_node

1 minute

instantaneous_output_kbps

Output Flow

Instantaneous output traffic

≥ 0

KiB/s

1024(IEC)

dcs_cluster_redis_node

1 minute

rejected_connections

Rejected Connections

Number of connections that have exceeded maxclients and been rejected during the monitoring period

≥ 0

N/A

N/A

dcs_cluster_redis_node

1 minute

expired_keys

Expired Keys

Number of keys that have expired and been deleted during the monitoring period

≥ 0

N/A

N/A

dcs_cluster_redis_node

1 minute

evicted_keys

Evicted Keys

Number of keys that have been evicted and deleted during the monitoring period

≥ 0

N/A

N/A

dcs_cluster_redis_node

1 minute

pubsub_channels

PubSub Channels

Number of Pub/Sub channels

≥ 0

N/A

N/A

dcs_cluster_redis_node

1 minute

pubsub_patterns

PubSub Patterns

Number of Pub/Sub patterns

≥ 0

N/A

N/A

dcs_cluster_redis_node

1 minute

keyspace_hits_perc

Hit Rate

Ratio of the number of Redis cache hits to the number of lookups. Calculation: keyspace_hits/(keyspace_hits + keyspace_misses)

If no read command is performed within a monitoring period, the ratio is 0.

0–100

%

N/A

dcs_cluster_redis_node

1 minute

command_max_delay

Maximum Command Latency

Maximum latency of commands

≥ 0

ms

N/A

dcs_cluster_redis_node

1 minute

is_slow_log_exist

Slow Query Logs

Existence of slow query logs in the node

Slow queries caused by the MIGRATE, SLAVEOF, CONFIG, BGSAVE, and BGREWRITEAOF commands are not counted.
  • 1: yes
  • 0: no

N/A

N/A

dcs_cluster_redis_node

1 minute

keys

Keys

Number of keys in Redis

≥ 0

N/A

N/A

dcs_cluster_redis_node

1 minute

sadd

SADD

Number of SADD commands processed per second

0–500,000

Count/s

N/A

dcs_cluster_redis_node

1 minute

smembers

SMEMBERS

Number of SMEMBERS commands processed per second

0–500,000

Count/s

N/A

dcs_cluster_redis_node

1 minute

ms_repl_offset

Replication Gap

Data synchronization gap between the master and the replica

Supported only by the replica node of master/standby, read/write splitting, and cluster DCS Redis instances.

-

Byte

1024(IEC)

dcs_cluster_redis_node

1 minute

del

DEL

Number of DEL commands processed per second

0–500,000

Count/s

N/A

dcs_cluster_redis_node

1 minute

expire

EXPIRE

Number of EXPIRE commands processed per second

0–500,000

Count/s

N/A

dcs_cluster_redis_node

1 minute

get

GET

Number of GET commands processed per second

0–500,000

Count/s

N/A

dcs_cluster_redis_node

1 minute

hdel

HDEL

Number of HDEL commands processed per second

0–500,000

Count/s

N/A

dcs_cluster_redis_node

1 minute

hget

HGET

Number of HGET commands processed per second

0–500,000

Count/s

N/A

dcs_cluster_redis_node

1 minute

hmget

HMGET

Number of HMGET commands processed per second

0–500,000

Count/s

N/A

dcs_cluster_redis_node

1 minute

hmset

HMSET

Number of HMSET commands processed per second

Supported only by master/standby and cluster DCS Redis instances.

0–500,000

Count/s

N/A

dcs_cluster_redis_node

1 minute

hset

HSET

Number of HSET commands processed per second

0–500,000

Count/s

N/A

dcs_cluster_redis_node

1 minute

mget

MGET

Number of MGET commands processed per second

0–500,000

Count/s

N/A

dcs_cluster_redis_node

1 minute

mset

MSET

Number of MSET commands processed per second

0–500,000

Count/s

N/A

dcs_cluster_redis_node

1 minute

set

SET

Number of SET commands processed per second

0–500,000

Count/s

N/A

dcs_cluster_redis_node

1 minute

rx_controlled

Flow Control Times

Number of times that client requests are controlled in a period. This metric is incremented by 1 each time a client request is controlled.

If the value is greater than 0, the consumed bandwidth exceeds the upper limit and flow control is triggered on a node. The node suspends client commands temporarily.

≥ 0

Count

N/A

dcs_cluster_redis_node

1 minute

bandwidth_usage

Bandwidth Usage

Percentage of the used bandwidth to the maximum bandwidth limit

0–200

%

N/A

dcs_cluster_redis_node

1 minute

connections_usage

Connection Usage

Percentage of the current number of connections to the maximum allowed number of connections

0–100

%

N/A

dcs_cluster_redis_node

1 minute

command_max_rt

Maximum Latency

Maximum delay from when the node receives commands to when it responds

≥ 0

μs

N/A

dcs_cluster_redis_node

1 minute

command_avg_rt

Average Latency

Average delay from when the node receives commands to when it responds

≥ 0

μs

N/A

dcs_cluster_redis_node

1 minute

sync_full

Full Sync Times

Total number of full synchronizations since the Redis Server last started

≥ 0

N/A

N/A

dcs_cluster_redis_node

1 minute

slow_log_counts

Slow Queries

Number of times that slow queries occur within a monitoring period

≥ 0

count

N/A

dcs_cluster_redis_node

1 minute

scan

SCAN

Number of SCAN operations per second

0–500,000

Count/s

N/A

dcs_cluster_redis_node

1 minute

setex

SETEX

Number of SETEX operations per second

0–500,000

Count/s

N/A

dcs_cluster_redis_node

1 minute

Proxy Metrics

  • These metrics are supported by Proxy Cluster and read/write splitting instances.
  • Dimensions lists the metric dimensions.
Table 4 Proxy metrics of Proxy Cluster or read/write splitting DCS Redis 4.0/5.0/6.0 instances

Metric ID

Metric Name

Metric Description

Value Range

Unit

Conversion Rule

Dimension

Monitoring Period (Raw Data)

node_status

Proxy Status

Indication of whether the proxy is normal.
  • 0: Normal
  • 1: Abnormal

N/A

N/A

dcs_cluster_proxy2_node

1 minute

cpu_usage

Maximum CPU usage

The monitored object's maximum CPU usage among multiple sampling values in a monitoring period

0–100

%

N/A

dcs_cluster_proxy2_node

1 minute

cpu_avg_usage

Average CPU Usage

The monitored object's average CPU usage of multiple sampling values in a monitoring period

0–100

%

N/A

dcs_cluster_proxy2_node

1 minute

memory_usage

Memory Usage

Memory consumed by the monitored object

0–100

%

N/A

dcs_cluster_proxy2_node

1 minute

connected_clients

Connected Clients

Number of connected clients. Includes connections established for system monitoring, configuration synchronization, and services.

≥ 0

N/A

N/A

dcs_cluster_proxy2_node

1 minute

instantaneous_ops

Ops per Second

Number of commands processed per second

≥ 0

N/A

N/A

dcs_cluster_proxy2_node

1 minute

instantaneous_input_kbps

Input Flow

Instantaneous input traffic

≥ 0

KiB/s

1024(IEC)

dcs_cluster_proxy2_node

1 minute

instantaneous_output_kbps

Output Flow

Instantaneous output traffic

≥ 0

KiB/s

1024(IEC)

dcs_cluster_proxy2_node

1 minute

total_net_input_bytes

Network Input Bytes

Number of bytes received during the monitoring period

≥ 0

byte

1024(IEC)

dcs_cluster_proxy2_node

1 minute

total_net_output_bytes

Network Output Bytes

Number of bytes sent during the monitoring period

≥ 0

byte

1024(IEC)

dcs_cluster_proxy2_node

1 minute

connections_usage

Connection Usage

Percentage of the current number of connections to the maximum allowed number of connections

0–100

%

N/A

dcs_cluster_proxy2_node

1 minute

command_max_rt

Maximum Latency

Maximum delay from when the node receives commands to when it responds

≥ 0

μs

N/A

dcs_cluster_proxy2_node

1 minute

command_avg_rt

Average Latency

Average delay from when the node receives commands to when it responds

≥ 0

μs

N/A

dcs_cluster_proxy2_node

1 minute

If a monitored object has multiple dimensions, the dimensional level of specific metrics is required when you use APIs to query the metrics.

For example, to query the maximum CPU usage (cpu_usage) of a DCS data node, its dimension is "dcs_instance_id,dcs_cluster_redis_node", indicating that dcs_instance_id is numbered 0 and dcs_cluster_redis_node is numbered 1.
  • To query a single metric by calling the API, the dcs_cluster_redis_node dimension is used as follows: 
    dim.0=dcs_instance_id,ca3c18f7-xxxx-xxxx-xxxx-76140724f2e4&dim.1=dcs_cluster_redis_node,b6258192xxxxxxxxx380a60c01f6

    ca3c18f7-xxxx-xxxx-xxxx-76140724f2e4 and b6258192xxxxxxxxx380a60c01f6 are the values of dcs_instance_id and dcs_cluster_redis_node, respectively. For details about how to obtain the values, see the obtaining guide in the Dimension table.

  • To batch query metrics by calling the API, the dcs_cluster_redis_node dimension is used as follows: 
    "dimensions": [ 
                 { 
                     "name": "dcs_instance_id", 
                     "value": "ca3c18f7-xxxx-xxxx-xxxx-76140724f2e4"    
                 }, 
                 { 
                     "name": "dcs_cluster_redis_node", 
                     "value": "b6258192xxxxxxxxx380a60c01f6" 
                 } 
             ]

    ca3c18f7-xxxx-xxxx-xxxx-76140724f2e4 and b6258192xxxxxxxxx380a60c01f6 are the values of dcs_instance_id and dcs_cluster_redis_node, respectively. For details about how to obtain the values, see the obtaining guide in the Dimension table.

Dimensions

Key

Value

dcs_instance_id

Redis instance ID, for example, ca3c18f7-xxxx-xxxx-xxxx-76140724f2e4.

To obtain the value, call the Querying All DCS Instances API and extract the value from the response parameter instance_id.

dcs_cluster_redis_node

Data node ID, for example, b6258192xxxxxxxxx380a60c01f6.

To obtain the value, call the Querying Instance Nodes API and extract the value from the response parameter logical_node_id.

dcs_cluster_proxy2_node

Proxy ID, for example, ff8080819axxxxxxxxb97ba16ae4.

To obtain the value, call the Querying Instance Nodes API and extract the value from the response parameter logical_node_id.

Related Documents