monitoring: CVE-Alarme aggregiert pro Image + Receiver-Robustheit (gitops#51)

Entscheidung sorb 2026-08-01 (Option 1 aus #51): Alarme als
count by (target, severity) statt pro CVE (~58 Serien statt ~1200),
CVE-Details bleiben im Dashboard (trivy_vuln_info unveraendert).
Receiver: inkrementelles save_state nach jedem Alarm, 1s-Sende-Drossel
(Synapse rc_message), recent_resolved-Dedup gegen doppelte Fallback-Haken
bei Batch-Retries, Teilfehler -> 502 liefert nur den Rest nach.
Stumm-Route + Null-Receiver entfernt - Zustellung wieder scharf.
promtool/amtool/py_compile gruen. UNGETESTET bis Deploy (Uebergabe-Issue).

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 16:11:25 +02:00
co-authored by Claude Fable 5
parent 0bd77e28d8
commit ff87cb25fb
3 changed files with 41 additions and 35 deletions
+3 -16
View File
@@ -5,25 +5,12 @@ route:
group_wait: 1m
group_interval: 5m
repeat_interval: 4h
routes:
# VORUEBERGEHEND STUMM (gitops#47): Die CVE-Regeln feuern eine Alarm-
# Instanz pro CVE pro Image. Messung an 3 der 29 Images ergab 27 CRITICAL
# und 126 HIGH -> hochgerechnet ~260 CRITICAL / ~1200 HIGH. matrix-alerts.py
# schickt eine Nachricht pro Alarm; Synapse rate-limitet nach ~10, die
# Exception verhindert save_state(), der Receiver antwortet 502 und
# Alertmanager wiederholt die komplette Gruppe -> Dauerschleife.
# Regeln laufen bewusst weiter (Zahlen sichtbar in Prometheus/Grafana),
# nur die Zustellung haengt hier. Wieder scharf schalten: diese route raus,
# sobald der Receiver Sammelnachrichten kann bzw. die Regeln aggregieren.
- matchers:
- room="security"
receiver: "null"
# CVE-Alarme laufen seit gitops#51 aggregiert (eine Instanz pro Image statt
# pro CVE) und wieder ueber den Matrix-Receiver — Historie des Alarm-Sturms
# und der Stummschaltung: gitops#51 / AAR 2026-08-01.
receivers:
- name: matrix
webhook_configs:
- url: http://matrix-alerts:8080/alert
send_resolved: true
# Schwarzes Loch fuer stummgeschaltete Routen - bewusst ohne Konfiguration.
- name: "null"
+21 -7
View File
@@ -12,6 +12,7 @@
import html
import json
import os
import time
import urllib.parse
import urllib.request
import uuid
@@ -30,15 +31,18 @@ ROOM_MAP = {k[len("MATRIX_ROOM_"):].lower(): v
try:
with open(STATE_FILE) as f:
state = json.load(f) # fingerprint -> {"event_id":..., "text":...}
_persisted = json.load(f)
state = _persisted.get("alerts", _persisted) # alt: flaches Format
recent_resolved = _persisted.get("recent_resolved", [])
except Exception:
state = {}
recent_resolved = []
def save_state():
try:
with open(STATE_FILE, "w") as f:
json.dump(state, f)
json.dump({"alerts": state, "recent_resolved": recent_resolved[-200:]}, f)
except Exception as e:
print(f"state save failed: {e}", flush=True)
@@ -74,28 +78,38 @@ class Handler(BaseHTTPRequestHandler):
if self.path != "/alert":
self.send_response(404); self.end_headers(); return
data = json.loads(self.rfile.read(int(self.headers.get("Content-Length", 0))))
try:
failed = 0
for a in data.get("alerts", []):
name = a.get("labels", {}).get("alertname", "?")
summary = a.get("annotations", {}).get("summary", "")
fp = a.get("fingerprint", "")
room = ROOM_MAP.get(a.get("labels", {}).get("room", ""), ROOM_ID)
try:
if a.get("status") == "firing":
text = f"\U0001F534 [firing] {name}: {summary}"
if fp in state: # re-notify desselben Alarms -> keine Doppelnachricht
if fp in state: # re-notify/Batch-Retry -> keine Doppelnachricht
continue
text = f"\U0001F534 [firing] {name}: {summary}"
event_id = send_text(text, room)
if fp and event_id:
state[fp] = {"event_id": event_id, "text": text, "room": room}
else:
if fp in recent_resolved: # Batch-Retry -> Fallback nicht doppeln
continue
known = state.pop(fp, None)
if known:
edit_resolved(known["event_id"], known["text"], known.get("room", ROOM_ID))
else:
send_text(f"✅ [resolved] {name}: {summary}", room)
save_state()
recent_resolved.append(fp)
save_state() # inkrementell: Teilfortschritt uebersteht Fehler/Retry
time.sleep(1) # Synapse-Ratelimit (rc_message) nicht reizen
except Exception as e:
self.log_message("matrix send failed: %s", e)
failed += 1
self.log_message("matrix send failed (%s): %s", name, e)
save_state()
time.sleep(2)
if failed:
# Alertmanager wiederholt den Batch; Dedup liefert nur den Rest nach
self.send_response(502); self.end_headers(); return
self.send_response(200); self.end_headers()
+11 -6
View File
@@ -58,21 +58,26 @@ groups:
# Alle Regeln routen per room-Label in den Security-Raum.
- name: axion-cve
rules:
- alert: TrivyCriticalVuln
expr: trivy_vuln_info{severity="CRITICAL"} == 1
# Aggregiert pro Image statt pro CVE (gitops#51, Entscheidung sorb
# 2026-08-01): ~1000 Einzelalarme waeren Rauschen; die menschlich
# handhabbare Einheit ist "Image X hat N kritische CVEs". Die
# CVE-Details (Pflichtfelder) liegen im Grafana-Dashboard cve-overview,
# die trivy_vuln_info-Einzelserien bleiben dafuer erhalten.
- alert: TrivyCriticalVulns
expr: count by (target, target_type, host) (trivy_vuln_info{severity="CRITICAL"}) > 0
labels:
severity: critical
room: security
annotations:
summary: "{{ $labels.cve }} (CRITICAL) in {{ $labels.target }} [{{ $labels.target_type }}@{{ $labels.host }}] — Paket {{ $labels.pkg }} {{ $labels.installed }}{{ if $labels.fixed_version }}, Fix: {{ $labels.fixed_version }}{{ else }}, noch kein Fix{{ end }} — https://nvd.nist.gov/vuln/detail/{{ $labels.cve }}"
- alert: TrivyHighVuln
expr: trivy_vuln_info{severity="HIGH"} == 1
summary: "{{ $labels.target }} [{{ $labels.target_type }}@{{ $labels.host }}]: {{ $value }} CRITICAL-CVEs — Details: https://selendis.axion1337.de/d/cve-overview"
- alert: TrivyHighVulns
expr: count by (target, target_type, host) (trivy_vuln_info{severity="HIGH"}) > 0
for: 24h
labels:
severity: warning
room: security
annotations:
summary: "{{ $labels.cve }} (HIGH) in {{ $labels.target }} [{{ $labels.target_type }}@{{ $labels.host }}] — Paket {{ $labels.pkg }} {{ $labels.installed }}{{ if $labels.fixed_version }}, Fix: {{ $labels.fixed_version }}{{ else }}, noch kein Fix{{ end }} — https://nvd.nist.gov/vuln/detail/{{ $labels.cve }}"
summary: "{{ $labels.target }} [{{ $labels.target_type }}@{{ $labels.host }}]: {{ $value }} HIGH-CVEs (seit 24h offen) — Details: https://selendis.axion1337.de/d/cve-overview"
- alert: TrivyScanStale
expr: time() - trivy_last_scan_timestamp > 172800
labels: