Redis Deploy & Monitor

Pigsty is also a universal application runtime that could be used for deploying & monitoring other databases and applications, such as Redis.

It takes two steps to deploy a Redis cluster:

1. Declare it
2. Execute playbook

Define Redis

• Redis Config Entries

E-R Model

The Redis entity conceptual model is almost identical to [PostgreSQL]c-pgsql.md#ER-Model), and also includes the concepts of Cluster and Instance. Note that the concept of Cluster here does not refer to the clusters in Redis’ native clustering scheme.

The core difference is that Redis is typically deployed in a single multi-instance deployment, with many Redis instances typically deployed on a single physical/virtual machine node to take advantage of multi-core CPUs. therefore, how Redis instances are defined is slightly different from PGSQL.

In Pigsty-managed Redis, the nodes are fully subordinate to the cluster, i.e. it is not currently allowed to deploy two Redis instances from different clusters on a single node, but this does not prevent you from deploying multiple independent Redis instances on a single node.

Redis Identity

The identity parameters are the information that must be provided when defining a Redis cluster and include.

• redis_cluster identifies the name of the Redis cluster, configured at the cluster level, and serves as the top-level namespace for cluster resources.
• redis_node identifies the serial number of the node in the cluster
• redis_instances is a JSON object with the Key as the instance port number and the Value as a JSON object containing the instance-specific configuration

Redis Cluster Definition

Given below are three condensed definitions of Redis clusters, including.

• A 1-node, 3-instance Redis Sentinel cluster redis-sentinel
• A 2-node, 12-instance Redis Cluster redis-cluster
• A 1-node, one-master-two-slave Redis Standalone cluster redis-standalone

You need to assign a unique port number to the Redis instance on the node.

#----------------------------------#
# sentinel example                 #
#----------------------------------#
redis-sentinel:
  hosts:
    10.10.10.10:
      redis_node: 1
      redis_instances:  { 6001 : {} ,6002 : {} , 6003 : {} }
  vars:
    redis_cluster: redis-sentinel
    redis_mode: sentinel
    redis_max_memory: 128MB

#----------------------------------#
# cluster example                  #
#----------------------------------#
redis-cluster:
  hosts:
    10.10.10.11:
      redis_node: 1
      redis_instances: { 6501 : {} ,6502 : {} ,6503 : {} ,6504 : {} ,6505 : {} ,6506 : {} }
    10.10.10.12:
      redis_node: 2
      redis_instances: { 6501 : {} ,6502 : {} ,6503 : {} ,6504 : {} ,6505 : {} ,6506 : {} }
  vars:
    redis_cluster: redis-cluster        # name of this redis 'cluster'
    redis_mode: cluster                 # standalone,cluster,sentinel
    redis_max_memory: 64MB              # max memory used by each redis instance
    redis_mem_policy: allkeys-lru       # memory eviction policy

#----------------------------------#
# standalone example               #
#----------------------------------#
redis-standalone:
  hosts:
    10.10.10.13:
      redis_node: 1
      redis_instances:
        6501: {}
        6502: { replica_of: '10.10.10.13 6501' }
        6503: { replica_of: '10.10.10.13 6501' }
  vars:
    redis_cluster: redis-standalone     # name of this redis 'cluster'
    redis_mode: standalone              # standalone,cluster,sentinel
    redis_max_memory: 64MB              # max memory used by each redis instance

Create Cluster

Deployment script

Create a Redis instance/cluster using the script redis.yml

./redis.yml -l redis-sentinel
./redis.yml -l redis-cluster
./redis.yml -l redis-standalone

Notes

Although this is not the recommended behavior, you can deploy PostgreSQL with Redis in a mixed deployment to make the most of machine resources.

The redis.yml script will deploy the Redis Monitor Exporter on the machine at the same time, including redis_exporter and node_exporter (optional)

During this process, the machine’s node_exporter will be redeployed if it exists.

By default, Prometheus will use the “multi-target crawl” mode, using the Redis Exporter on port 9121 on the node to crawl all Redis instances on that node.

Redis Monitoring

Pigsty currently provides 3 Redis monitoring panels as part of a standalone monitoring application REDIS, which are.

• Redis Overview: provides a global overview of Redis across the entire environment
• Redis Cluster: focuses on monitoring information for a single Redis business cluster
• Redis Instance: provides detailed monitoring information about a single Redis instance

You can use the included redis-benchmark test

CAVEAT

Pigsty v1.5.1 supports whole-cluster Redis deployment and monitoring, and the tagged redis.yml/redis-remove.yml playbooks also target one instance with -e redis_port=<port>.

Offline, scale-up, scale-down, and single-instance management features will be provided in subsequent versions gradually.