monitoring: CVE-Pipeline v1 (gitops#47) - Scanner, Exporter, Regeln, Routing, Dashboard

- cve-scan: Trivy-Loop ueber die 29 real deployten Images (Cluster-Inventur
  2026-08-01 + Prod-Web-Image); 24h-Intervall, Fehler einzelner Images
  blockieren nicht
- cve-exporter: Stdlib-Exporter mit first_seen-State (Zeitstrahl), Schema
  trivy_vuln_info/_count/_first_seen/_last_scan gemaess Pflichtfeldern
- 3 Alertregeln (CRITICAL sofort, HIGH mit 24h-Daempfung, Scan-Frische) -
  promtool SUCCESS 9 rules; alle mit room=security
- matrix-alerts: Label-basiertes Raum-Routing (MATRIX_ROOM_<NAME>), Edits
  landen im richtigen Raum via State
- Grafana-Dashboard cve-overview: Severity-Stats, CVE-Tabelle mit
  NVD-Link/Fix-Version/first-seen, Zeitstrahl, Verlauf
UNGETESTET bis zum Deploy auf CFGMON (compose up -d).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PKhFj1S3UdD6xL2fbWPeYj
This commit is contained in:
Thore Cimbal
2026-08-01 13:42:07 +02:00
co-authored by Claude Fable 5
parent f45e01219b
commit b6007c50fd
8 changed files with 300 additions and 10 deletions
+98
View File
@@ -0,0 +1,98 @@
#!/usr/bin/env python3
# Trivy-JSON -> Prometheus-Metriken (gitops#47). Stdlib-only, Machart wie die
# uebrigen Monitoring-Helfer. Liest die Reports des cve-scan-Sidecars und
# serviert /metrics; merkt sich je (CVE, Target) den Erstfund (first_seen),
# damit der geforderte Zeitstrahl (erstmals gesehen / geschlossen) abbildbar
# ist - "geschlossen" = Serie verschwindet, resolved kommt via Alertmanager.
#
# Schema (Pflichtfelder-Vorgabe sorb: CVE-ID, Mitigation, Zeitstrahl, Ort, Typ):
# trivy_vuln_info{cve,severity,target,target_type,host,pkg,installed,fixed_version} 1
# (nur HIGH/CRITICAL als Einzelserien - Kardinalitaet)
# trivy_vuln_count{target,target_type,host,severity} <n> (alle Severities)
# trivy_vuln_first_seen_timestamp{cve,target} <unix>
# trivy_last_scan_timestamp{target,target_type,host} <unix>
import json
import os
import time
from http.server import BaseHTTPRequestHandler, ThreadingHTTPServer
RESULTS = os.environ.get("RESULTS_DIR", "/results")
STATE_FILE = os.environ.get("STATE_FILE", "/state/first-seen.json")
# Ort: Container-Images laufen (bis auf Weiteres) alle auf dem MATRIX-Host;
# Host-rootfs-Scans (Ausbaustufe) bringen ihren Hostnamen im Dateinamen mit.
DEFAULT_HOST = os.environ.get("DEFAULT_HOST", "matrix")
try:
with open(STATE_FILE) as f:
first_seen = json.load(f) # "cve|target" -> unix-ts
except Exception:
first_seen = {}
def esc(v):
return str(v).replace("\\", "\\\\").replace('"', '\\"').replace("\n", " ")
def collect():
lines = []
now = int(time.time())
seen_keys = set()
for fn in sorted(os.listdir(RESULTS)):
if not fn.endswith(".json"):
continue
path = os.path.join(RESULTS, fn)
try:
rep = json.load(open(path))
except Exception:
continue
target = rep.get("ArtifactName", fn[:-5])
ttype = "host" if rep.get("ArtifactType") in ("filesystem", "rootfs") else "image"
host = fn.split("__host__")[1].split(".json")[0] if "__host__" in fn else DEFAULT_HOST
base = f'target="{esc(target)}",target_type="{ttype}",host="{esc(host)}"'
lines.append(f'trivy_last_scan_timestamp{{{base}}} {int(os.path.getmtime(path))}')
counts = {}
for res in rep.get("Results") or []:
for v in res.get("Vulnerabilities") or []:
sev = v.get("Severity", "UNKNOWN")
counts[sev] = counts.get(sev, 0) + 1
if sev in ("HIGH", "CRITICAL"):
cve = v.get("VulnerabilityID", "?")
key = f"{cve}|{target}"
if key not in first_seen:
first_seen[key] = now
seen_keys.add(key)
lines.append(
'trivy_vuln_info{cve="%s",severity="%s",%s,pkg="%s",installed="%s",fixed_version="%s"} 1'
% (esc(cve), sev, base, esc(v.get("PkgName", "?")),
esc(v.get("InstalledVersion", "?")), esc(v.get("FixedVersion", ""))))
lines.append(f'trivy_vuln_first_seen_timestamp{{cve="{esc(cve)}",target="{esc(target)}"}} {first_seen[key]}')
for sev, n in sorted(counts.items()):
lines.append(f'trivy_vuln_count{{{base},severity="{sev}"}} {n}')
# State kompakt halten: verschwundene Findings raus (= "geschlossen")
for k in list(first_seen):
if k not in seen_keys:
del first_seen[k]
try:
os.makedirs(os.path.dirname(STATE_FILE), exist_ok=True)
with open(STATE_FILE, "w") as f:
json.dump(first_seen, f)
except Exception as e:
print(f"state save failed: {e}", flush=True)
return "\n".join(lines) + "\n"
class Handler(BaseHTTPRequestHandler):
def do_GET(self):
if self.path != "/metrics":
self.send_response(404); self.end_headers(); return
body = collect().encode()
self.send_response(200)
self.send_header("Content-Type", "text/plain; version=0.0.4")
self.end_headers()
self.wfile.write(body)
def log_message(self, fmt, *args):
pass
ThreadingHTTPServer(("0.0.0.0", 9101), Handler).serve_forever()
+29
View File
@@ -0,0 +1,29 @@
clamav/clamav:1.5.3
coturn/coturn:latest
docker.io/grafana/alloy:v1.7.5
docker.io/library/haproxy:3.2-alpine
docker.io/library/postgres:17.9-bookworm
docker.io/livekit/livekit-server:v1.10.0
docker.io/postgres:17-alpine
docker.io/prometheuscommunity/postgres-exporter:v0.18.1
ghcr.io/element-hq/ess-helm/matrix-tools:0.7.3
ghcr.io/element-hq/matrix-authentication-service:1.15.0
ghcr.io/fluxcd/helm-controller:v1.5.3
ghcr.io/fluxcd/kustomize-controller:v1.8.3
ghcr.io/fluxcd/notification-controller:v1.8.3
ghcr.io/fluxcd/source-controller:v1.8.2
ghcr.io/goauthentik/server:2026.2.3
gnuxie/draupnir:v3.1.0
nginx:1.26-alpine
oci.element.io/element-admin:0.1.11
oci.element.io/lk-jwt-service:0.4.2
oci.element.io/synapse:v1.151.0-ess.1
quay.io/jetstack/cert-manager-cainjector:v1.14.0
quay.io/jetstack/cert-manager-controller:v1.14.0
quay.io/jetstack/cert-manager-webhook:v1.14.0
quay.io/prometheus-operator/prometheus-config-reloader:v0.81.0
registry.k8s.io/kube-state-metrics/kube-state-metrics:v2.15.0
rohana.axion1337.de/sorb/axion-backup:v2
rohana.axion1337.de/sorb/axion-secret-rotation:v1
rohana.axion1337.de/sorb/clamav-http-scanner:v1.0.0
rohana.axion1337.de/sorb/threadnet-web:v0.3.0
+23
View File
@@ -0,0 +1,23 @@
#!/bin/sh
# Trivy-Scan-Schleife (gitops#47): scannt alle 24h die Liste der real deployten
# Images (cve/images.txt, aus dem Cluster inventarisiert) und legt die
# JSON-Reports fuer den cve-exporter ab. Kein Abbruch bei Einzelfehlern -
# ein nicht mehr pullbares Image darf den Rest nicht verhindern (der Exporter
# alarmiert ueber trivy_last_scan_timestamp, wenn ein Report veraltet).
set -u
RESULTS=/results
IMAGES=/config/images.txt
INTERVAL="${SCAN_INTERVAL_SECONDS:-86400}"
while true; do
while IFS= read -r img; do
case "$img" in ""|\#*) continue;; esac
safe=$(echo "$img" | tr '/:@' '___')
echo "scan: $img"
trivy image --scanners vuln --format json --output "$RESULTS/$safe.json.tmp" "$img" \
&& mv "$RESULTS/$safe.json.tmp" "$RESULTS/$safe.json" \
|| echo "FEHLER bei $img (Report bleibt auf altem Stand)"
done < "$IMAGES"
echo "runde fertig, schlafe ${INTERVAL}s"
sleep "$INTERVAL"
done