Redis
Overview
Learn how to use Redis for high-performance caching, real-time data streaming, and inter-service communication in the Sravz platform. This section covers Redis Timeseries for storing and querying time-series financial data.
What You’ll Learn
- Redis Timeseries: Purpose-built module for time-series data
- Real-time Data Storage: Store stock quotes with timestamps
- Efficient Queries: Range queries, aggregations, and downsampling
- High Performance: Sub-millisecond latency for reads/writes
- Compaction: Automatic data retention and downsampling
- Integration: Use Redis with C++, Go, Node.js, and Python backends
Key Technologies
- Redis 7.x: In-memory data store
- RedisTimeSeries: Time-series database module
- Redis Streams: Message broker for event streaming
- Redis Pub/Sub: Real-time messaging
- RedisJSON: Store and query JSON documents
- Redis Search: Full-text search and indexing
Documentation Index
- Redis Timeseries Introduction - Introduction to Redis Timeseries module for financial time-series data
Use Cases
Time-Series Data
- Stock Quotes: Store real-time tick data
- OHLCV Bars: 1-minute, 5-minute, hourly, daily bars
- Trading Signals: Strategy signals with timestamps
- Performance Metrics: System performance monitoring
Caching
- S3 Presigned URLs: Cache expensive URL generation
- API Responses: Cache external API calls
- Database Queries: Cache frequent MongoDB queries
- Session Data: User session storage
Streaming
- Redis Streams: Event sourcing and message queues
- Pub/Sub: Real-time notifications
- Inter-service Communication: Connect C++, Go, Node.js services
- WebSocket Broadcasting: Push updates to clients
Redis Timeseries Features
Data Storage
TS.ADD ticker:AAPL 1638360000000 150.25 LABELS symbol AAPL type quote
Range Queries
TS.RANGE ticker:AAPL 1638360000000 1638446400000 AGGREGATION avg 3600000
Compaction Rules
- Automatically downsample high-resolution data
- Multiple retention policies (1min → 1hour → 1day)
- Save storage while preserving historical data
Aggregations
- avg: Average price over time window
- min/max: Price range
- sum: Total volume
- count: Number of samples
- first/last: Opening/closing prices
Architecture
Data Flow
Market Data → Backend (C++/Go) → Redis Timeseries → Query Layer → UI
↓
Downsampling & Retention
Integration Points
- C++ Backend: Boost Redis client
- Go Backend: go-redis library
- Node.js: ioredis client
- Python: redis-py
- Rust: redis-rs
Performance Benefits
Speed
- In-Memory: Sub-millisecond latency
- Optimized Storage: Compressed time-series format
- Fast Queries: Native time-range operations
- Batch Operations: Pipeline multiple commands
Scalability
- Redis Cluster: Horizontal scaling
- Replication: High availability
- Persistence: AOF and RDB snapshots
- Compression: Efficient storage format
Common Patterns
Real-time Ticker Plant
WebSocket → Parse → Redis Timeseries → [Streams] → Consumers
Cache-Aside Pattern
Request → Check Redis → Hit? Return : Query DB → Cache → Return
Event Sourcing
Events → Redis Streams → Consumers → Process → Update State
Best Practices
Time-Series
- Use Labels: Tag series with metadata (symbol, type)
- Retention Policies: Define appropriate TTLs
- Compaction Rules: Downsample old data
- Duplicate Policy: Handle duplicate timestamps
Caching
- TTL: Set expiration for cache entries
- Eviction Policy: Use LRU for memory limits
- Cache Warming: Pre-populate frequent queries
- Invalidation: Clear cache on data updates
Streaming
- Consumer Groups: Distribute work across instances
- ACKs: Acknowledge processed messages
- Dead Letter Queue: Handle failed messages
- Monitoring: Track lag and throughput
Getting Started
- Read Redis Timeseries Introduction
- Install Redis with TimeSeries module
- Store sample stock quotes
- Query and aggregate data
- Set up compaction rules
Advanced Topics
- Redis Cluster: Sharding for large datasets
- Redis Streams: Event-driven architectures
- RedisJSON: Store complex documents
- RedisSearch: Full-text indexing
- RedisGraph: Graph database capabilities