Redis Caching Strategy

Relevant source files

The following files were used as context for generating this wiki page:

• backend/src/config/redis.ts
• backend/src/infrastructure/cache/RedisCache.ts
• infrastructure/aws/cdk/lib/atlasclean-stack.ts
• infrastructure/terraform/main.tf
• docker-compose.yml
• README.md

The Redis Caching Strategy provides a high-performance data storage layer for session coordination, asynchronous work orchestration, and application-level caching. Redis serves as a volatile data store that reduces latency for frequently accessed information and supports the platform's AI-powered scheduling and dispatch optimization.

Sources: README.md:14, README.md:21

Infrastructure Architecture

The project implements Redis using distinct configurations for local development, staging, and production environments. In cloud environments, the system utilizes AWS ElastiCache for Redis, deployed within private, isolated subnets for security.

Cloud Deployment (AWS)

The infrastructure is defined via both AWS CDK and Terraform. In production, the system deploys a Replication Group with automatic failover and Multi-AZ support.

Configuration Details

Option	Development (Docker)	Staging (AWS)	Production (AWS)
Node Type	Container (7-alpine)	cache.t4g.micro	cache.t4g.small
Cluster Count	1	1	2 (Primary + Replica)
Encryption	Disabled	In-transit & At-rest	In-transit & At-rest
Failover	Manual	Disabled	Automatic Enabled

Sources: infrastructure/aws/cdk/lib/atlasclean-stack.ts:133-146, infrastructure/terraform/main.tf:302-318, docker-compose.yml:32-38

Implementation Logic

Connection Management

The redisClient utilizes ioredis to manage connections to the Redis instance. It is configured with lazyConnect and enableAutoPipelining to optimize performance. The system maintains a singleton instance of the client in non-production environments to prevent connection leaks during hot reloads.

export const redisClient =
  globalForRedis.redis ??
  new Redis(config.redisUrl, {
    lazyConnect: true,
    maxRetriesPerRequest: 3,
    enableAutoPipelining: true,
  });

Sources: backend/src/config/redis.ts:8-14

Cache Service Interface

The RedisCache class abstracts the underlying client, providing a simplified interface for standard cache operations. It automatically handles the serialization and deserialization of objects using JSON.stringify and JSON.parse.

Method	Action	Parameters
get<T>	Retrieves and parses a value from the cache	key: string
set	Serializes and stores a value	key: string, value: unknown, ttlSeconds?: number
del	Removes an entry by key	key: string

Sources: backend/src/infrastructure/cache/RedisCache.ts:3-22

Data Flow for Cache Access

The following sequence diagram illustrates how the application interacts with the caching layer when performing a set operation with an expiration time.

Security and Connectivity

• Security Groups: Access is restricted via EC2 security groups. Only resources within the ServiceSecurityGroup (ECS Tasks) are permitted to connect to the CacheSecurityGroup on port 6379.
• Networking: Redis is deployed into isolated subnets (SubnetType.PRIVATE_ISOLATED) with no direct internet access.
• Authentication: Connection strings are provided via the REDIS_URL environment variable, which includes credentials in local development or points to the primary endpoint address in AWS.

Sources: infrastructure/aws/cdk/lib/atlasclean-stack.ts:101-104, infrastructure/terraform/main.tf:217-226, backend/src/config/redis.ts:9

Conclusion

The Redis Caching Strategy provides the AtlasCleanAI platform with a robust, scalable caching layer. By utilizing ioredis for connection management and a specialized RedisCache utility for data operations, the system achieves low-latency access to session and operational data while maintaining high security through AWS network isolation and encryption.