feature : connection pooling

[HatemMn]

Idea :

In a service using Redis, connection pooling (via r2d2) optimizes performance by maintaining a fixed set of reusable connections, eliminating the overhead of repeatedly establishing new ones.
For a machine with N CPUs, setting the pool size to N is a balance: it maximizes parallel request handling without under-utilizing resources (smaller pools) or introducing contention (larger pools). This approach ensures efficient concurrency while minimizing Redis client initialization costs, as recommended in redis-rs best practices. ( source doc )

Notes :

Already implemented in some branch, but not taken to prod

ref :

• https://github.com/redis-rs/redis-rs?tab=readme-ov-file#connection-pooling
• https://usavps.com/blog/112017/