Hardware recommendations and scaling guidelines for every deployment model — from single-node proof of concept to multi-data-center enterprise rollouts.
| Deployment Tier | Database Systems | CPU Cores | Memory | Storage | Use Case |
|---|---|---|---|---|---|
| Proof of Concept | Up to 3 | 2 cores | 4 GB | 50 GB | Evaluation and feature demonstration |
| Small Critical | Up to 25 | 4-8 cores | 16 GB | 500 GB | Production HA for critical systems |
| Classical | Up to 500 | 8-16 cores | 32-64 GB | 2+ TB | Enterprise with distributed gathering |
| Large / Hybrid | 1000+ | 16+ cores | 64+ GB | 10+ TB | Multi-data-center, multi-cloud |
Scalefield Secure scales horizontally. Each monitored database gets its own Gather instance (lightweight, runs as a sidecar or standalone pod). The Secure Engine scales by adding replicas. Storage grows linearly but is mitigated by Parquet compression (8-12x). For a typical OLTP database generating 10,000 audit events per minute, expect approximately 50-100 MB of compressed Parquet data per day.
Parquet's columnar format with dictionary and run-length encoding achieves 8-12x compression. A terabyte of raw audit logs becomes less than 100 GB of archived Parquet data.
Configure retention periods per regulatory framework. Automated lifecycle management moves data between hot and cold tiers, and purges expired records on schedule.
Start with local filesystem for PoC, scale to Ceph or S3 for production. Migration between backends is supported without data loss or downtime.
Need help sizing your deployment? Talk to our team.