Integrate a Redis/Valkey instance-Cloud Monitor(CMS)-阿里云帮助中心

This topic describes how to integrate a Redis/Valkey instance with Cloud Monitor 2.0. The integration component is based on the Redis adapter of aio-exporter and collects metrics such as performance, memory, client connections, keyspace hits and evictions, persistence, and master-slave replication. It supports Redis 5.x and later (version 6.2 or later is required for full functionality), is compatible with Valkey 7.x, 8.x, and 9.x, and can be deployed in Container Service for Kubernetes (ACK) clusters and on Elastic Compute Service (ECS) instances in a virtual private cloud (VPC) environment. After integration, Redis instance entities and a built-in dashboard are automatically generated.

Requirements

Before you start the integration, make sure that the target instance meets the following requirements. Otherwise, no data may be collected or some metrics may be missing.

Condition	Requirement
Version	Redis 5.x or later. Full functionality requires Redis 6.2 or later, and latency histograms require Redis 7.0 or later. An ACL username requires Redis 6.0 or later. Valkey 7.x, 8.x, and 9.x are compatible.
Deployment type	Self-managed Redis or Valkey. A cloud-managed instance can also be integrated if it meets the account and network conditions.
Authentication	(Recommended) For Redis 6.0 and later, create a dedicated ACL monitoring user. If only `requirepass` is configured, enter the password and leave the username blank. Enter the username and the password in separate fields. Do not include them in the address.
Network connectivity	In the CS environment, the target Redis instance must be reachable within the network of the selected ACK cluster. In the ECS (VPC) environment, the vSwitch CIDR block of the fully managed collector must be able to access the target ECS instance.

Prerequisites

Managed Service for Prometheus is activated.
CS environment — The target ACK cluster is integrated with Cloud Monitor 2.0 and is in the Running state. The target Redis service is accessible from the cluster network.
ECS (VPC) environment — The target Redis service is deployed on an ECS instance.

Billing

Billing is based solely on the volume of metric data generated in Billing overview.

Integration methods

Step 1: Prepare a monitoring account (if needed)

For Redis 6.0 or later, we recommend that you create a separate Access Control List (ACL) user for monitoring. If you use only requirepass for password authentication, you can provide a password and leave the username empty.

ACL SETUSER exporter on ><password> -@all +@connection +memory -readonly +strlen +config|get +xinfo +pfcount -quit +zcard +type +xlen -readwrite -command +client -wait +scard +llen +hlen +arcount +get +eval +slowlog +cluster|info +cluster|slots +cluster|nodes -hello -echo +info +latency +scan -reset -auth -asking

The ACL user must have read-only permissions for commands such as INFO, CONFIG GET, CLIENT, SLOWLOG, LATENCY, and MEMORY. In Sentinel scenarios, permissions for commands such as +sentinel|masters are also required.

Step 2: Select an integration environment

This plugin supports two access environments: CS and ECS (VPC). Select the appropriate environment based on where your Redis instance is deployed:

The Exporter is installed in the specified ACK cluster and reuses the existing data collector in the cluster to collect metrics.
The monitored object is the ACK cluster.
Network requirements: The target Redis service must be reachable from the cluster network. For example, it can be a Service within the cluster or an internal VPC address that the cluster can access.

ECS (VPC)

A fully managed data collector is deployed based on the VPC information of the ECS instance where the Redis service is deployed.
No processes are installed on the ECS instance.
The monitored object is the ECS instance.
Network requirements: The vSwitch CIDR block for the managed data collector must have access to the target ECS instance. You must configure security groups, ACLs, and other rules to allow access.

Environment	Exporter deployment location	Process installed on workload host?	Monitored object	Network connectivity requirements
CS	A specified ACK cluster	No. The Exporter is deployed in the cluster and does not run on the application's host node.	ACK cluster	Redis must be reachable from the cluster network.
ECS (VPC)	Fully managed data collector deployed by the system based on the VPC	No. No processes are installed on the ECS instance.	ECS instance	The vSwitch CIDR block of the managed agent must have access to the target ECS instance.

Step 3: Configure the integration

Log on to the Cloud Monitor console, and select a workspace. In the navigation pane on the left, click Integration Center. Then, click Redis to start the integration.
Select CS or ECS (VPC) as the access environment, and then select an Observation object from the drop-down list.

Complete the following configuration and click OK:

Configuration Item	Required	Description
Redis Endpoint	Yes	For a CS environment, enter the in-cluster endpoint. For an ECS (VPC) environment, enter the internal or public endpoint. Do not include the username or password in the endpoint.
Redis Service Port	Yes	Default: 6379
Cloud Vendor/Deployment Source	No	This value is written to the `provider` tag. For self-hosted deployments, enter `self-hosted/redis`.
Network Connectivity Test	—	Available only for CS environments.
Redis Username	No	The username for a Redis 6.0+ ACL user. If you leave this empty, the system uses the `default` user (password only).
Redis Logon Password	No	The password for the monitoring account.
Metric Collection Interval (seconds)	Yes	Default: 15
Custom Tags	No	Additional tags to add to metrics.
Use Public Registry Address	No	Disabled by default. Enable this option if the internal image repository is inaccessible.

After the integration is complete, you can view the integrated entities and their observable data in Entity Explorer.

Data usage

Entities

After you integrate Redis, Cloud Monitor periodically discovers Redis instances from the labels of metrics such as redis_up and creates middleware entities.

Dimension	Description
Domain	middleware
Entity type	middleware.redis
Entity filter condition	`.entity with(domain='middleware', type='middleware.redis')`
Observed object (bound entity)	For CS environments, the object is `acs.ack.cluster`. For ECS (VPC) environments, the object is `acs.ecs.instance`.

Entity fields are derived from collected labels and are not related to the number of installations.

Field	Description
instance_id	Instance identity. This field is derived from the `target_uri` label in the `host:port` format.
provider	Deployment source, such as `self-hosted/redis`.
port	Service port.
region_id	Region.
env_type	Deployment environment type (CS or ECS).
cluster_id	ID of the cluster to which the instance belongs.

You can retrieve instances using these fields in the Entity Explorer or Middleware Insight application.

Built-in dashboards

Dashboard name	Access point	Description
Instance Overview	The Instance Overview tab on the entity details page.	UModel dashboard. Variables are automatically bound based on the instance context.
Redis Overview	Grafana dashboard list.	Grafana dashboard. Available in both Chinese and English.

Core metrics

The following is a complete list of the 64 metrics for the middleware.metric.redis entity dataset. You can use these metrics directly on the entity details page, in dashboards, and in alerts.

Metric name	Type	Unit	Display name	Description
redis_up	gauge	-	Redis availability	Indicates whether the Redis/Valkey instance is reachable and data collection is successful. A value of 1 indicates that the instance is available. A value of 0 indicates that the instance is unavailable.
redis_uptime	gauge	Seconds	Uptime	The number of seconds that the Redis service process has been running since it started.
redis_clients	gauge	-	Client connections	The number of client connections to Redis. This does not include connections from replicas.
redis_blocked	gauge	-	Blocked clients	The number of clients waiting due to blocking commands, such as BLPOP, BRPOP, and BZPOP.
redis_slaves	gauge	-	Number of secondary nodes	The number of connected replicas (secondary nodes).
redis_commands_s	counter	Commands/s	Commands per second	The total number of commands processed per second.
redis_hit_rate	counter	%	Cache hit rate	The keyspace hit rate, calculated as hits / (hits + misses) and expressed as a percentage.
redis_evicted_s	counter	Keys/s	Evicted keys per second	The rate at which keys are evicted because the maxmemory limit is reached.
redis_memory_usage	gauge	%	Memory usage	The percentage of used memory relative to the maxmemory limit.
redis_used	gauge	Bytes	Used memory	The total number of bytes allocated by Redis for data (used_memory).
redis_max	gauge	Bytes	Maximum memory	The configured maxmemory limit in bytes. A value of 0 indicates no limit.
redis_rss	gauge	Bytes	RSS memory	Resident Set Size (RSS) memory. The number of bytes of resident memory that the Redis process occupies from the operating system's perspective.
redis_peak	gauge	Bytes	Peak memory	The peak value of used_memory since the instance started.
redis_dataset	gauge	Bytes	Dataset memory	The number of bytes occupied by the dataset. This does not include metadata overhead.
redis_overhead	gauge	Bytes	Memory overhead	The number of bytes occupied by overhead, such as Redis metadata and internal structures.
redis_fragmentation	gauge	-	Memory fragmentation ratio	The ratio of RSS to used_memory. A value significantly higher than 1.5 may indicate memory fragmentation.
redis_allocator_frag	gauge	-	Allocator fragmentation ratio	The memory allocator fragmentation ratio reported by Redis.
redis_tracking	gauge	-	Monitoring the client count	The number of clients that have client-side caching tracking enabled.
redis_received_s	counter	Times per second	New connections per second	The rate at which new client connections are accepted.
redis_rejected_s	counter	Count per second	Rejected connections per second	The rate at which connections are rejected because the maxclients limit is reached.
redis_input	counter	Bytes	Inbound traffic	The number of bytes received from the network per second.
redis_output	counter	Bytes	Outbound traffic	The number of bytes sent to the network per second.
redis_max_input	gauge	Bytes	Maximum input buffer	The maximum number of bytes in the client input buffer recently.
redis_max_output	gauge	Bytes	Maximum output buffer	The maximum number of bytes in the client output buffer recently.
redis_top_command_calls_sec	counter	-	Top command calls	The call rate of the top 10 commands, aggregated by command type.
redis_command_duration_sec_by_cmd	counter	-	Top command duration	The execution duration rate (in seconds) of the top 10 commands, aggregated by command type.
redis_failed	counter	-	Failed commands	The call rate of failed commands, aggregated by command type.
redis_rejected	counter	-	Rejected commands	The call rate of rejected commands, such as out-of-memory (OOM) errors, aggregated by command type.
redis_sys	counter	-	System CPU	The system mode CPU time consumed per second by the main Redis process.
redis_user	counter	-	User CPU	The user mode CPU time consumed per second by the main Redis process.
redis_sys_children	counter	-	Child process system CPU	The system mode CPU time consumed per second by background child processes, such as bgsave.
redis_user_children	counter	-	Child process user CPU	The user mode CPU time consumed per second by background child processes.
redis_keys_per_db	gauge	-	Keys per DB	The number of keys in each logical database.
redis_non_expiring	gauge	-	Non-expiring keys	The number of keys that do not have a time-to-live (TTL) set.
redis_expiring	gauge	-	Expiring keys	The number of keys that have a TTL set.
redis_hits	counter	-	Keyspace hits	The rate of successful key lookups (hits).
redis_misses	counter	-	Keyspace misses	The rate of failed key lookups (misses).
redis_expired	counter	-	Expired keys	The rate at which keys are deleted due to expiration.
redis_avg_ttl_per_db	gauge	Seconds	Average TTL per DB	The average TTL (in seconds) of keys with an expiration time in each database.
redis_rdb_bgsave_ok	gauge	-	RDB background save status	Indicates whether the last background RDB save was successful. 1 indicates success, and 0 indicates failure.
redis_rdb_bgsave_in_progress	gauge	-	RDB background save in progress	Indicates whether a background RDB save is in progress. 1 indicates yes, and 0 indicates no.
redis_aof_enabled	gauge	-	AOF enabled	Indicates whether Append-Only File (AOF) persistence is enabled. 1 indicates yes, and 0 indicates no.
redis_aof_bgrewrite_ok	gauge	-	AOF rewrite status	Indicates whether the last background AOF rewrite was successful. 1 indicates success, and 0 indicates failure.
redis_aof_write_ok	gauge	-	AOF write status	Indicates whether the last AOF write was successful. 1 indicates success, and 0 indicates failure.
redis_loading	gauge	-	Loading data	Indicates whether data is being loaded from a dump file. 1 indicates yes, and 0 indicates no.
redis_lag	gauge	Seconds	Replica lag	The replication delay (in seconds) of a connected replica relative to the primary node.
redis_master_offset	gauge	-	Primary replication offset	The current replication offset of the primary node.
redis_slave_offset	gauge	-	Secondary replication offset	The current replication offset of this replica.
redis_master_link_up	gauge	-	Primary-secondary link status	Indicates whether the link between the replica and the primary node is active. 1 indicates active, and 0 indicates disconnected.
redis_master_last_io_ago	gauge	Seconds	Time since last primary IO	The number of seconds since the last interaction with the primary node, from the replica's perspective.
redis_changes_since_last_save	gauge	-	Changes since last save	The number of changes since the last successful RDB save.
redis_last_bgsave_duration	gauge	-	Last BGSAVE duration	The duration (in seconds) of the last background RDB save.
redis_aof_size	gauge	Bytes	Current AOF size	The current size of the AOF file in bytes.
redis_channels	gauge	-	Pub/Sub channels	The number of active Publish/Subscribe (Pub/Sub) channels.
redis_patterns	gauge	-	Pub/Sub Pattern Count	The number of active Pub/Sub pattern subscriptions.
redis_pubsub_clients	gauge	-	Pub/Sub clients	The number of clients in Pub/Sub mode.
redis_length	gauge	-	Slow log entries	The number of entries in the current slow query log.
redis_last_slow_duration	gauge	Seconds	Last slow command duration	The execution duration (in seconds) of the most recent command in the slow query log.
redis_latency_spikes	gauge	Seconds	Latency spike duration	The duration of recent latency spikes reported by Redis latency monitoring.
redis_cycles_s	gauge	Cycles/s	Event loop cycles per second	The instantaneous number of event loop cycles per second.
redis_duration_usec	gauge	-	Event loop duration	The instantaneous duration of the event loop in microseconds.
redis_error_replies_s	counter	Replies/s	Error replies per second	The rate of error replies sent to clients.
redis_error_replies	counter	-	Error replies by type	The rate of error replies, aggregated by error type.
redis_last_scrape_error	gauge	-	Last scrape error	Indicates whether the last redis_exporter scrape failed. 1 indicates failure, and 0 indicates success.

In addition to the preceding MetricSet aggregation metrics, all raw metrics (prefixed with redis_*) collected by the exporter are written to Prometheus storage. You can query these metrics directly using Prometheus Query Language (PromQL).

Common query examples

# Availability
redis_up

# Memory usage
redis_memory_used_bytes / clamp_min(redis_memory_max_bytes, 1) * 100

# Hit rate
rate(redis_keyspace_hits_total[5m])
  / clamp_min(rate(redis_keyspace_hits_total[5m]) + rate(redis_keyspace_misses_total[5m]), 1)

Alert templates

This integration component does not have built-in alert rules. You can configure alerts in the Alert Center based on the metrics described previously. For example, you can set alerts for scenarios where the redis_up availability metric is not 1, or when the number of connections or message accumulation exceeds a threshold.

Storage policies

After you integrate Redis, Cloud Monitor creates a corresponding storage instance to store observable data.

Data type	Default storage location	Notes
Metrics	Prometheus instance: RegionShare:{{workspaceName}}:{{regionId}}	Associated with the `middleware.metric.redis` dataset.

Manage or remove the integration

To modify or cancel the integration, go to the Integration Management page of Integration Center, find the corresponding integration policy, and then click Edit or Delete. When you delete the integration policy, the Exporter and the collection configuration are removed together. Historical metrics that have already been written are retained based on the lifecycle policy of the storage.