mirror of
https://github.com/marcredhat/SIEM-toolkit-patched
synced 2026-06-08 20:37:12 +00:00
Mick
c182d837ee
Dockerized SecOps toolkit with: - Coverage Map: STAR rule vs SDL parser field coverage analysis - Ingest Dashboard: PowerQuery-powered event volume and source breakdown - Onboarding Assistant: AI-guided log source onboarding with Claude - Parser management via SDL MCP integration Stack: FastAPI + PostgreSQL backend, nginx-served HTML frontend, Docker Compose. PowerQuery runs via Scalyr XDR API (SDL_XDR_URL + SDL_LOG_READ_KEY). Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
102 lines
3.8 KiB
Python
102 lines
3.8 KiB
Python
from datetime import datetime, timedelta
|
|
from fastapi import APIRouter, Query, HTTPException
|
|
from pydantic import BaseModel
|
|
from services import s1_client
|
|
|
|
router = APIRouter()
|
|
|
|
|
|
def _date_range(days: int) -> tuple[str, str]:
|
|
now = datetime.utcnow()
|
|
return (
|
|
(now - timedelta(days=days)).strftime("%Y-%m-%dT%H:%M:%S.000Z"),
|
|
now.strftime("%Y-%m-%dT%H:%M:%S.000Z"),
|
|
)
|
|
|
|
|
|
@router.get("/top-sources")
|
|
async def get_top_sources(days: int = Query(7, ge=1, le=90)):
|
|
"""Top log sources by event count over the given period."""
|
|
from_dt, to_dt = _date_range(days)
|
|
query = "| group events=count() by dataSource.name | sort -events | limit 25"
|
|
try:
|
|
result = await s1_client.run_powerquery(query, from_dt, to_dt)
|
|
except Exception as e:
|
|
raise HTTPException(502, f"PowerQuery error: {e}")
|
|
return {"period_days": days, "data": result.get("events", [])}
|
|
|
|
|
|
@router.get("/by-event-type")
|
|
async def get_by_event_type(days: int = Query(7, ge=1, le=90)):
|
|
"""Event counts grouped by source and event type."""
|
|
from_dt, to_dt = _date_range(days)
|
|
query = "| group events=count() by dataSource.name, event.type | sort -events | limit 100"
|
|
try:
|
|
result = await s1_client.run_powerquery(query, from_dt, to_dt)
|
|
except Exception as e:
|
|
raise HTTPException(502, f"PowerQuery error: {e}")
|
|
return {"period_days": days, "data": result.get("events", [])}
|
|
|
|
|
|
@router.get("/daily-volume")
|
|
async def get_daily_volume(days: int = Query(7, ge=1, le=14)):
|
|
"""Total event count per day."""
|
|
import asyncio
|
|
results = []
|
|
points = min(days, 7)
|
|
for i in range(points):
|
|
day_from = (datetime.utcnow() - timedelta(days=i + 1)).strftime("%Y-%m-%dT00:00:00.000Z")
|
|
day_to = (datetime.utcnow() - timedelta(days=i)).strftime("%Y-%m-%dT00:00:00.000Z")
|
|
label = (datetime.utcnow() - timedelta(days=i + 1)).strftime("%Y-%m-%d")
|
|
try:
|
|
result = await s1_client.run_powerquery("| group total=count()", day_from, day_to)
|
|
events_list = result.get("events") if isinstance(result, dict) else []
|
|
count = events_list[0].get("total", 0) if isinstance(events_list, list) and events_list else 0
|
|
except Exception:
|
|
count = 0
|
|
results.append({"date": label, "events": count})
|
|
if i < points - 1:
|
|
await asyncio.sleep(3)
|
|
return list(reversed(results))
|
|
|
|
|
|
class FilterRule(BaseModel):
|
|
source: str = ""
|
|
event_type: str = ""
|
|
days: int = 7
|
|
gb_per_million_events: float = 0.5
|
|
|
|
|
|
@router.post("/simulate-filter")
|
|
async def simulate_filter(rule: FilterRule):
|
|
"""Estimate how many events and GB would be eliminated by an exclusion filter."""
|
|
from_dt, to_dt = _date_range(rule.days)
|
|
|
|
clauses = []
|
|
if rule.source:
|
|
clauses.append(f'src.name = "{rule.source}"')
|
|
if rule.event_type:
|
|
clauses.append(f'event.type = "{rule.event_type}"')
|
|
|
|
filter_expr = " AND ".join(clauses) if clauses else "true"
|
|
query = f"| filter {filter_expr} | count() as events"
|
|
|
|
try:
|
|
result = await s1_client.run_powerquery(query, from_dt, to_dt)
|
|
events = (result.get("events") or [{}])[0].get("events", 0) if isinstance(result.get("events"), list) else 0
|
|
except Exception as e:
|
|
raise HTTPException(502, f"PowerQuery error: {e}")
|
|
|
|
estimated_gb = round(events / 1_000_000 * rule.gb_per_million_events, 3)
|
|
monthly_events = int(events / rule.days * 30)
|
|
monthly_gb = round(monthly_events / 1_000_000 * rule.gb_per_million_events, 2)
|
|
|
|
return {
|
|
"period_days": rule.days,
|
|
"matched_events": events,
|
|
"estimated_gb_period": estimated_gb,
|
|
"projected_monthly_events": monthly_events,
|
|
"projected_monthly_gb": monthly_gb,
|
|
"filter": {"source": rule.source, "event_type": rule.event_type},
|
|
}
|