NetCrunch Databases
NetCrunch uses several specialized databases to efficiently manage configuration, event history, real-time status, long-term metrics, and documentation. Each database is optimized for its specific role in the monitoring platform, allowing NetCrunch to scale from hundreds to thousands of nodes while ensuring high performance.
Overview
The architecture separates data by function:
- Atlas holds configuration and topology.
- Event Log captures historical actions and alerts.
- Status DB maintains real-time state.
- Trend DB stores long-term performance metrics.
- DocDB manages auxiliary documentation like notes and inventory.
This separation improves scalability, simplifies backup strategies, and enables faster access to different data types.
NetCrunch Atlas (XML, JSON)
NetCrunch Atlas stores all network objects, their state, monitoring configurations, maps, and dependencies. It is maintained in-memory for performance and periodically saved to disk as multiple XML (and optionally JSON) files.
- Type: In-memory with periodic disk save
- Format: XML (primary), JSON (optional export)
- Scope: Nodes, maps, monitoring packs, policies, groups
NetCrunch Event Log (SQL)
The event log stores all system events and action execution logs. It is stored in a SQL database.
- Type: SQL (on server)
- Engine: PostgreSQL
- Access: ODBC-compatible
- Use: Audit trail, historical analysis, compliance
-
Disk Requirements:
- SATA drives support a few GB
- Use SSD or RAID for > 5 GB of daily events
NetCrunch Status DB (NoSQL)
The Status DB is an in-memory hierarchical NoSQL database used to track the live state of monitored objects, triggers, and monitors.
- Type: In-memory NoSQL with persistent disk save
- Scope: Triggers, monitor states, external sensor overrides
- Realtime Updates: Supports dynamic changes from external sensors or agents
NetCrunch Performance Trend Database (NoSQL)
This is a specialized append-only NoSQL database optimized for storing and querying large volumes of performance metrics.
- Storage Model: Daily journaling files, compressed and aggregated
- Scale: Capable of handling 250+ million records per day
- Use: Long-term trend analysis (weekly/monthly/yearly)
- Optimized For: High write throughput and efficient time-series access
NetCrunch DocDB (NoSQL)
DocDB is a document-oriented database built on RocksDB. It is used for storing metadata and documentation linked to nodes.
- Engine: RocksDB (embedded)
-
Contents:
- Node notes
- Task assignments
- Device configurations
- Hardware/software/hotfixes inventory history
Summary
Each NetCrunch database serves a distinct role. This architecture allows NetCrunch to balance real-time responsiveness with historical analysis and robust configuration management. It also enables efficient use of resources across environments of different sizes.