Introduction

Redis (“REmote DIctionary Service”) is an open-source key-value database server.

Transclude of row-based-and-column-based-database#memory-vs-disk-based-databases

Queues, Pub/sub, and Streams

Feature	Redis Queue aka Redis List	Redis Pub/Sub	Redis Streams
Purpose	Task/job scheduling and background processing.	Real-time messaging between publishers and subscribers.	Real-time data processing with persistence and consumer groups.
Data Model	FIFO (First In, First Out) queue.	Publish-subscribe messaging system.	Append-only log of key-value pairs.
Persistence	No inherent persistence; tasks are lost if not processed before a crash.	No persistence; messages are lost if subscribers are offline.	Persistent; entries remain in the stream until explicitly trimmed.
Scalability	Can distribute jobs across multiple workers for horizontal scaling.	Scales well with a large number of publishers and subscribers.	Supports multiple consumers via consumer groups for distributed processing.
Reliability	Supports job retries, prioritization, and dependencies for reliable task execution.	Fire-and-forget; no guarantees on message delivery or retries.	Provides at-least-once delivery with acknowledgment mechanisms for reliability.
Use Cases	- Background job processing - Task scheduling - Distributed workloads - Job prioritization and dependencies.	- Real-time notifications - Chat applications - Live updates - Event broadcasting.	- IoT data ingestion - Real-time analytics - Message brokering - Logging/audit trails. - Activity feeds.
Consumer Model	Workers explicitly pull tasks from the queue (pull-based).	Subscribers passively receive messages when published (push-based).	Consumers actively read entries from the stream; consumer groups enable parallel processing (pull-based).
Message Retention	Tasks remain in the queue until processed or manually deleted.	Messages are ephemeral and only delivered to active subscribers at the time of publishing.	Messages persist in the stream until explicitly trimmed or expired based on retention policies.

Redis High Availability

High Availability (HA) in Redis ensures that the database remains operational even if some nodes fail. It is crucial for applications that require minimal downtime and data consistency.
08 - ESB & ASG > Scalability & High Availability

Redis provides two primary HA solutions:

Redis Sentinal – Provides automatic failover but does NOT support Database Sharding.
Redis Cluster – Provides both failover and data sharding for scalability.

Redis Sentinel Vs. Redis Cluster

Redis Sentinel vs Clustering | Baeldung

Feature	Redis Sentinel	Redis Cluster
Purpose	Automatic Failover	Failover + Data Sharding
Data Sharding	❌ No	✅ Yes
Replication	✅ Master-Replica	✅ Master-Replica
Failover Handling	✅ Yes	✅ Yes
Minimum Nodes	At least 3 Sentinels	At least 3 Masters + Replicas
Best For	Failover & HA only	Scalability + HA

Nimendra's Notes 🪴

Explorer

Redis

TOC

Introduction

Queues, Pub/sub, and Streams

Redis High Availability

Redis Sentinel Vs. Redis Cluster

Dead Letter Queues

Fire and Forget Messaging

Redis Cluster

Redis Data Persistence

Redis Geospatial

Redis HSET, EXPIRE & TTL

Redis PubSub and Implementation

Redis Queue

Redis Reliable Queue Pattern

Redis Sentinal

Redis Serialization Protocol(RESP)

Redis Sorted Sets

Redis Streams