Usage Tracking
VARICON uses a post-processing, report-driven usage model.
Unlike traditional observability tools, it does not track raw traffic or logs.
Core Principle
Usage is derived from processed anomalies, not raw data.
Where Usage Happens
Usage is measured at a single point:
AFTER enrichment
BEFORE projectionPipeline Context
Capture → Transport → Extraction → Normalization → Fingerprinting
→ Enrichment → [USAGE COUNTED HERE] → Projection → OutputWhy This Position
This ensures: • only valuable work is billed • no cost for detection or transport • no dependence on raw API traffic
Unit of Measurement
1 enriched anomaly = 1 unitData Flow
Step-by-step
- Newman produces report
- Extraction engine parses anomalies
- Canonical anomalies created
- Enrichment applied
- Usage counter increments
- Projection filters output
- Artifacts generated
Usage Record Model
{
"project_id": "proj_abc123",
"run_id": "run_xyz789",
"timestamp": 1773956107,
"anomaly_count": 1200,
"enriched": true
}Aggregation
Usage is aggregated: • per CI run • per project • per billing window
Determinism
Usage tracking must be: • deterministic • reproducible • idempotent
Example
Same input → same anomaly count → same billing
Idempotency Strategy
To prevent double counting: • each run has a unique run_id • repeated processing of same report does not increase usage
Offline Mode
If VARICON runs without enrichment: • anomalies are still processed • reports are generated • usage = 0
Fail-Open Behavior
If enrichment fails: • anomalies are returned • usage is not incremented • pipeline continues
Security Boundary
Usage tracking operates only on: • counts • metadata
It never processes: • raw request bodies • raw response bodies • sensitive payloads
Comparison to Traditional Systems
| System | Billing Unit |
|---|---|
| Datadog | logs |
| Sentry | events |
| Postman | seats |
| VARICON | enriched anomalies |
Architectural Advantage
VARICON avoids: • log ingestion pipelines • streaming infrastructure • high storage costs
Summary
VARICON measures value, not volume.
Usage grows with: • system complexity • test coverage • anomaly density
Key Insight
The more your system drifts from expected behavior, the more valuable VARICON becomes.