MMT API Best Practices
Rules for building reliable, performant applications with the MMT market data API.
Connection Management
- WebSocket Lifecycle: connect/disconnect flow, heartbeat, connection states, cleanup
- Authentication: API key via header (REST) vs query param (WS), server-side proxy for browser apps
- Reconnection Strategy: exponential backoff, resubscription, state recovery
Rate Limiting
- Weight Management: weight costs per endpoint, budget allocation, monitoring
- Tier Awareness: Strict/Standard/Premium limits, feature gating
- Burst Strategy: burst bucket mechanics, pre-flight checks, graceful degradation
Data Handling
- CBOR Optimization: when and how to use CBOR, bandwidth savings, decoding
- Symbol Normalization: base/quote format, /markets discovery, inverse contracts
- Timeframe Rules: validation rules, calendar timeframes, UTC alignment
- Type Handling: public type field maps for all data types
WebSocket
- Subscription Management: sub/unsub protocol, limits, deduplication
- Range Requests: historical data via WS, from/to params
- Channel Selection: channel guide and when to use each
REST API
- Endpoint Selection: endpoint decision tree, parameter reference
- Caching Strategy: cache /markets, TTLs, ETag patterns
- Timeframe Alignment: UTC boundary alignment, candle close semantics
Multi-Exchange Aggregation
- Multi-Exchange Queries: colon-separated syntax, supported endpoints
- Cost Optimization: weight multipliers, tier restrictions
Error Handling
- Error Patterns: error code reference, structured format, recovery
- Retry Strategy: retryable vs non-retryable, backoff, circuit breaker
Performance
- Memory & Batching: batch fetches, stream processing, avoid unbounded buffers