I see a lot of teams treating vulnerability scans and threat intel as separate silos. That's a missed opportunity. I've built a pipeline that correlates Tenable Nessus findings with ThreatConnect threat intelligence to prioritize what to patch *now* based on active adversary tactics.
The core idea is simple: map CVEs from Tenable to threat indicators in ThreatConnect. The execution requires some glue. I avoid vendor-specific agents and use OpenTelemetry for data collection wherever possible.
Here's the high-level flow:
1. **Extract:** Pull CVE data from Tenable.io via its API on a scheduled basis (e.g., daily). I use a lightweight Python script for this.
2. **Transform:** Flatten the Nessus data, focusing on host IP, plugin ID, and most importantly, the associated CVE list.
3. **Correlate:** Query ThreatConnect's API for indicators (like Malware, Tools, Attack Patterns) that are linked to those CVEs. I filter for indicators with a higher confidence rating (e.g., > 50%).
4. **Enrich & Emit:** The output is a new dataset—hosts with vulnerabilities that have known, active threat associations. This gets sent as OTLP logs and metrics to my observability backend.
```yaml
# Snippet from the config for the correlation service
correlation_engine:
tenable_scan_interval: "24h"
threatconnect_min_confidence: 50
output:
otlp_endpoint: "http://otel-collector:4318"
metrics:
- name: "vuln.with_active_threat"
description: "Count of vulnerabilities linked to high-confidence threat indicators"
unit: "1"
```
The result is a dashboard in Grafana that doesn't just show me vulnerable systems. It highlights systems where the vulnerabilities are being actively exploited in the wild. My Prometheus alerts are now tuned to fire on `vuln.with_active_threat > 0` instead of generic CVE counts.
The pitfall? ThreatConnect's API can be sluggish on large CVE batches, so you need to implement caching and polite throttling. Also, this setup is only as good as your threat intel community's data in TC—if they aren't tagging indicators with CVEs, you'll get false negatives.
-ninja
Observability is not monitoring