Streamlining infrastructure monitoring
Industry:
Technology
Outcome:
Tech stack simplification from nine tools to one
Tencent integrated SurrealDB to unify its infrastructure monitoring stack. By leveraging native graph capabilities, versioned data access, and built-in analytics, the team replaced a fragmented toolchain with a scalable, real-time monitoring platform.
Key challenges
Fragmented toolchain
Monitoring data lived across multiple databases and stream processors, making it hard to correlate events and maintain consistency.
High operational costs
Analysts had to choose between storage and compute engines, increasing the platform team's maintenance burden.
Limited real-time analytics
Existing systems lacked native graph views or recursive traversal for tracing relationships and state transitions.
Massive log volumes
Scaling ingestion and queries for millions of time-series logs per day proved challenging.
Solutions
Graph-first modelling
Modelled infrastructure components as nodes with edge-based relationships to enable recursive traversal and event causality.
Versioned time-series data
Stored snapshots and change logs with complex record IDs for efficient time-travel queries and temporal filtering.
Pre-computed views
Persisted aggregated metrics into table views to reduce dashboard latency and enforce policies.
Live monitoring
Implemented LIVE SELECT on log tables for real-time alerts and diagnostics without polling.
Results
Unified data layer
1 SYSTEM
SurrealDB replaced separate time-series, document, and graph stores to simplify the stack.
Consistent low-latency access
↓ LATENCY
Delivered both historical snapshots and real-time streams from a single system.
Scalable log ingestion
↑ LOGS/DAY
Compound record keys and range filters handled millions of logs per day without schema changes.
Real-time alerts
INSTANT
Live queries powered alerting and UI refreshes for faster operational response.
