feat(audit): add comprehensive audit logging with SIEM integration

- Add AuditLog model with append-only design (timestamp, user, action, resource, IP, details, severity)
- Add audit_service.py with record/query helpers and SIEM forwarding (Syslog RFC 5424, HTTP/webhook)
- Add /api/audit-logs REST endpoints with filtering and pagination
- Add /admin/audit-logs viewer UI with real-time filters
- Add SIEM config settings (syslog, HTTP for Splunk HEC/Logstash/Grafana Loki)
- Add Alembic migration 027_add_audit_logs
- Add navigation link in admin menu
- Add comprehensive tests (20 tests covering model, service, SIEM, API, view)

Co-authored-by: christianlouis <361235+christianlouis@users.noreply.github.com>
This commit is contained in:
copilot-swe-agent[bot]
2026-03-09 23:18:18 +00:00
parent 7685b787f2
commit 35db9f88de
14 changed files with 1188 additions and 0 deletions
+2
View File
@@ -8,6 +8,7 @@ from fastapi import APIRouter
from app.api.admin_users import router as admin_users_router
from app.api.api_tokens import router as api_tokens_router
from app.api.audit_logs import router as audit_logs_router
from app.api.azure import router as azure_router
from app.api.backup import router as backup_router
from app.api.billing import router as billing_router
@@ -82,3 +83,4 @@ router.include_router(imap_accounts_router)
router.include_router(integrations_router)
router.include_router(notifications_router)
router.include_router(scheduled_jobs_router)
router.include_router(audit_logs_router)
+115
View File
@@ -0,0 +1,115 @@
"""
Audit log REST API endpoints.
Provides read-only access to the comprehensive audit log for admin users.
Events are append-only — there are no update or delete endpoints.
"""
import logging
from datetime import datetime
from typing import Any
from fastapi import APIRouter, Depends, Query, Request
from sqlalchemy.orm import Session
from app.auth import require_login
from app.database import get_db
from app.utils.audit_service import count_events, query_events
logger = logging.getLogger(__name__)
router = APIRouter()
@router.get("/audit-logs")
@require_login
async def list_audit_logs(
request: Request,
db: Session = Depends(get_db),
action: str | None = Query(None, description="Filter by action (exact match)"),
user: str | None = Query(None, description="Filter by username"),
resource_type: str | None = Query(None, description="Filter by resource type"),
severity: str | None = Query(None, description="Filter by severity level"),
since: datetime | None = Query(None, description="Only events at or after this ISO-8601 timestamp"),
until: datetime | None = Query(None, description="Only events at or before this ISO-8601 timestamp"),
limit: int = Query(50, ge=1, le=500, description="Max rows to return"),
offset: int = Query(0, ge=0, description="Rows to skip for pagination"),
) -> dict[str, Any]:
"""Return audit log entries with optional filtering and pagination.
Requires authentication. Returns events in reverse chronological order.
"""
entries = query_events(
db,
action=action,
user=user,
resource_type=resource_type,
severity=severity,
since=since,
until=until,
limit=limit,
offset=offset,
)
total = count_events(
db,
action=action,
user=user,
resource_type=resource_type,
severity=severity,
since=since,
until=until,
)
return {
"items": [_serialize(e) for e in entries],
"total": total,
"limit": limit,
"offset": offset,
}
@router.get("/audit-logs/actions")
@require_login
async def list_distinct_actions(
request: Request,
db: Session = Depends(get_db),
) -> list[str]:
"""Return the distinct action values present in the audit log."""
from app.models import AuditLog
rows = db.query(AuditLog.action).distinct().order_by(AuditLog.action).all()
return [r[0] for r in rows]
@router.get("/audit-logs/users")
@require_login
async def list_distinct_users(
request: Request,
db: Session = Depends(get_db),
) -> list[str]:
"""Return the distinct user values present in the audit log."""
from app.models import AuditLog
rows = db.query(AuditLog.user).distinct().order_by(AuditLog.user).all()
return [r[0] for r in rows]
# ------------------------------------------------------------------
# Helpers
# ------------------------------------------------------------------
def _serialize(entry) -> dict[str, Any]:
"""Convert an AuditLog row to a JSON-safe dict."""
import json as _json
return {
"id": entry.id,
"timestamp": entry.timestamp.isoformat() if entry.timestamp else None,
"user": entry.user,
"action": entry.action,
"resource_type": entry.resource_type,
"resource_id": entry.resource_id,
"ip_address": entry.ip_address,
"details": _json.loads(entry.details) if entry.details else None,
"severity": entry.severity,
}
+43
View File
@@ -849,6 +849,49 @@ class Settings(BaseSettings):
),
)
# SIEM / External Audit Log Forwarding
# Forward audit events to external SIEM systems for centralised monitoring.
audit_siem_enabled: bool = Field(
default=False,
description="Enable forwarding of audit events to an external SIEM system.",
)
audit_siem_transport: str = Field(
default="syslog",
description=(
"Transport used to forward audit events. "
"Options: 'syslog' (RFC 5424 over UDP/TCP), 'http' (JSON POST to a webhook URL, "
"compatible with Splunk HEC, Logstash HTTP input, Grafana Loki, etc.)."
),
)
audit_siem_syslog_host: str = Field(
default="localhost",
description="Hostname or IP of the syslog receiver.",
)
audit_siem_syslog_port: int = Field(
default=514,
description="Port of the syslog receiver.",
)
audit_siem_syslog_protocol: str = Field(
default="udp",
description="Protocol for syslog transport: 'udp' or 'tcp'.",
)
audit_siem_http_url: str = Field(
default="",
description=(
"HTTP endpoint URL for SIEM webhook delivery. "
"Supports Splunk HEC (https://splunk:8088/services/collector/event), "
"Logstash HTTP input, Grafana Loki push API, or any JSON-accepting endpoint."
),
)
audit_siem_http_token: str = Field(
default="",
description="Bearer / HEC token included in the Authorization header of SIEM HTTP requests.",
)
audit_siem_http_custom_headers: str = Field(
default="",
description="Comma-separated 'Key:Value' pairs of extra headers for SIEM HTTP requests.",
)
# UI / Appearance
ui_default_color_scheme: str = Field(
default="system",
+21
View File
@@ -146,6 +146,27 @@ class SettingsAuditLog(Base):
action = Column(String, nullable=False) # "update" or "delete"
class AuditLog(Base):
"""Comprehensive audit log for compliance tracking.
Records all significant actions: login/logout, document CRUD, settings
changes, and administrative operations. Rows are append-only; the API
and service layer never update or delete entries.
"""
__tablename__ = "audit_logs"
id = Column(Integer, primary_key=True, index=True)
timestamp = Column(DateTime(timezone=True), server_default=func.now(), nullable=False, index=True)
user = Column(String, nullable=False, index=True) # Username or "anonymous" / "system"
action = Column(String, nullable=False, index=True) # e.g. "login", "document.create", "settings.update"
resource_type = Column(String, nullable=True, index=True) # e.g. "document", "user", "settings"
resource_id = Column(String, nullable=True) # ID of the affected resource
ip_address = Column(String, nullable=True) # Client IP address
details = Column(Text, nullable=True) # JSON-encoded extra context
severity = Column(String(16), nullable=False, server_default="info") # info / warning / error / critical
class SavedSearch(Base):
"""User-defined saved search filters for quick access to frequently used filter combinations."""
+319
View File
@@ -0,0 +1,319 @@
"""
Comprehensive audit-event service for DocuElevate.
Provides helpers to **record** audit events (append-only database writes)
and to optionally **forward** them to external SIEM systems.
Supported SIEM transports:
* **Syslog** RFC 5424 structured-data messages over UDP or TCP.
* **HTTP** JSON POST payloads compatible with Splunk HEC, Logstash
HTTP input, Grafana Loki push API, and any generic webhook endpoint.
"""
import json
import logging
import socket
import threading
from datetime import datetime, timezone
from typing import Any
import httpx
from fastapi import Request
from sqlalchemy.orm import Session
from app.config import settings
from app.middleware.audit_log import get_client_ip, get_username
from app.models import AuditLog
logger = logging.getLogger(__name__)
# ---------------------------------------------------------------------------
# Public helpers
# ---------------------------------------------------------------------------
def record_event(
db: Session,
*,
action: str,
user: str = "system",
resource_type: str | None = None,
resource_id: str | None = None,
ip_address: str | None = None,
details: dict[str, Any] | None = None,
severity: str = "info",
) -> AuditLog:
"""Persist an audit event and optionally forward it to SIEM.
Args:
db: Active SQLAlchemy session.
action: Short action identifier (e.g. ``"login"``, ``"document.create"``).
user: Username performing the action.
resource_type: Category of the affected resource (``"document"``, ``"user"`` …).
resource_id: Identifier of the affected resource.
ip_address: Client IP address (``None`` when not applicable).
details: Arbitrary key/value context serialised as JSON.
severity: One of ``info``, ``warning``, ``error``, ``critical``.
Returns:
The newly created :class:`AuditLog` row.
"""
details_json = json.dumps(details, default=str) if details else None
entry = AuditLog(
user=user,
action=action,
resource_type=resource_type,
resource_id=str(resource_id) if resource_id is not None else None,
ip_address=ip_address,
details=details_json,
severity=severity,
)
db.add(entry)
db.commit()
db.refresh(entry)
# Fire-and-forget SIEM forwarding in a background thread so we never
# block the request path.
if settings.audit_siem_enabled:
payload = _build_siem_payload(entry)
thread = threading.Thread(target=_forward_to_siem, args=(payload,), daemon=True)
thread.start()
return entry
def record_event_from_request(
db: Session,
request: Request,
*,
action: str,
resource_type: str | None = None,
resource_id: str | None = None,
details: dict[str, Any] | None = None,
severity: str = "info",
) -> AuditLog:
"""Convenience wrapper that extracts user and IP from a :class:`Request`.
Args:
db: Active SQLAlchemy session.
request: The current HTTP request.
action: Short action identifier.
resource_type: Category of the affected resource.
resource_id: Identifier of the affected resource.
details: Arbitrary key/value context serialised as JSON.
severity: One of ``info``, ``warning``, ``error``, ``critical``.
Returns:
The newly created :class:`AuditLog` row.
"""
return record_event(
db,
action=action,
user=get_username(request),
resource_type=resource_type,
resource_id=resource_id,
ip_address=get_client_ip(request),
details=details,
severity=severity,
)
def query_events(
db: Session,
*,
action: str | None = None,
user: str | None = None,
resource_type: str | None = None,
severity: str | None = None,
since: datetime | None = None,
until: datetime | None = None,
limit: int = 200,
offset: int = 0,
) -> list[AuditLog]:
"""Query audit log entries with optional filtering.
Args:
db: Active SQLAlchemy session.
action: Filter by action string (exact match).
user: Filter by username (exact match).
resource_type: Filter by resource type (exact match).
severity: Filter by severity level (exact match).
since: Only events at or after this timestamp.
until: Only events at or before this timestamp.
limit: Maximum number of rows to return.
offset: Number of rows to skip (for pagination).
Returns:
List of :class:`AuditLog` rows ordered by *timestamp descending*.
"""
q = db.query(AuditLog)
if action:
q = q.filter(AuditLog.action == action)
if user:
q = q.filter(AuditLog.user == user)
if resource_type:
q = q.filter(AuditLog.resource_type == resource_type)
if severity:
q = q.filter(AuditLog.severity == severity)
if since:
q = q.filter(AuditLog.timestamp >= since)
if until:
q = q.filter(AuditLog.timestamp <= until)
return q.order_by(AuditLog.timestamp.desc()).offset(offset).limit(limit).all()
def count_events(
db: Session,
*,
action: str | None = None,
user: str | None = None,
resource_type: str | None = None,
severity: str | None = None,
since: datetime | None = None,
until: datetime | None = None,
) -> int:
"""Return the total count of events matching the given filters.
Args:
db: Active SQLAlchemy session.
action: Filter by action string.
user: Filter by username.
resource_type: Filter by resource type.
severity: Filter by severity level.
since: Only events at or after this timestamp.
until: Only events at or before this timestamp.
Returns:
Integer count.
"""
q = db.query(AuditLog)
if action:
q = q.filter(AuditLog.action == action)
if user:
q = q.filter(AuditLog.user == user)
if resource_type:
q = q.filter(AuditLog.resource_type == resource_type)
if severity:
q = q.filter(AuditLog.severity == severity)
if since:
q = q.filter(AuditLog.timestamp >= since)
if until:
q = q.filter(AuditLog.timestamp <= until)
return q.count()
# ---------------------------------------------------------------------------
# SIEM forwarding internals
# ---------------------------------------------------------------------------
_SYSLOG_FACILITY_LOCAL0 = 16
_SYSLOG_SEVERITY_MAP = {
"info": 6,
"warning": 4,
"error": 3,
"critical": 2,
}
def _build_siem_payload(entry: AuditLog) -> dict[str, Any]:
"""Convert an :class:`AuditLog` row into a plain dict for SIEM delivery."""
ts = entry.timestamp if entry.timestamp else datetime.now(timezone.utc)
return {
"id": entry.id,
"timestamp": ts.isoformat(),
"user": entry.user,
"action": entry.action,
"resource_type": entry.resource_type,
"resource_id": entry.resource_id,
"ip_address": entry.ip_address,
"details": entry.details,
"severity": entry.severity,
"source": "docuelevate",
}
def _forward_to_siem(payload: dict[str, Any]) -> None:
"""Route a SIEM payload to the configured transport."""
transport = settings.audit_siem_transport.lower()
try:
if transport == "syslog":
_send_syslog(payload)
elif transport == "http":
_send_http(payload)
else:
logger.warning("Unknown SIEM transport %r; skipping forwarding", transport)
except Exception:
logger.exception("Failed to forward audit event to SIEM (%s)", transport)
def _send_syslog(payload: dict[str, Any]) -> None:
"""Send a RFC 5424 syslog message to the configured receiver."""
severity_num = _SYSLOG_SEVERITY_MAP.get(payload.get("severity", "info"), 6)
priority = _SYSLOG_FACILITY_LOCAL0 * 8 + severity_num
ts = payload.get("timestamp", datetime.now(timezone.utc).isoformat())
hostname = socket.gethostname()
app_name = "docuelevate"
msg_id = payload.get("action", "-")
# Structured data (SD) element with key event fields.
sd = (
f'[docuelevate@0 user="{payload.get("user", "-")}" '
f'action="{payload.get("action", "-")}" '
f'resource_type="{payload.get("resource_type", "-")}" '
f'resource_id="{payload.get("resource_id", "-")}" '
f'ip="{payload.get("ip_address", "-")}"]'
)
message = json.dumps(payload, default=str)
syslog_msg = f"<{priority}>1 {ts} {hostname} {app_name} - {msg_id} {sd} {message}"
proto = settings.audit_siem_syslog_protocol.lower()
host = settings.audit_siem_syslog_host
port = settings.audit_siem_syslog_port
if proto == "tcp":
with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as sock:
sock.settimeout(5)
sock.connect((host, port))
sock.sendall(syslog_msg.encode("utf-8"))
else:
with socket.socket(socket.AF_INET, socket.SOCK_DGRAM) as sock:
sock.settimeout(5)
sock.sendto(syslog_msg.encode("utf-8"), (host, port))
logger.debug("Syslog audit event sent to %s:%s (%s)", host, port, proto)
def _send_http(payload: dict[str, Any]) -> None:
"""POST a JSON audit event to the configured HTTP endpoint."""
url = settings.audit_siem_http_url
if not url:
logger.warning("SIEM HTTP URL not configured; skipping HTTP forwarding")
return
headers: dict[str, str] = {"Content-Type": "application/json"}
token = settings.audit_siem_http_token
if token:
headers["Authorization"] = f"Bearer {token}"
# Parse custom headers (comma-separated "Key:Value" pairs).
raw_custom = settings.audit_siem_http_custom_headers
if raw_custom:
for raw_pair in raw_custom.split(","):
pair = raw_pair.strip()
if ":" in pair:
k, _, v = pair.partition(":")
headers[k.strip()] = v.strip()
# Wrap in Splunk HEC-style envelope when URL contains ``/services/collector``.
body: dict[str, Any]
if "/services/collector" in url:
body = {"event": payload, "sourcetype": "docuelevate:audit", "source": "docuelevate"}
else:
body = payload
with httpx.Client(timeout=10) as client:
resp = client.post(url, json=body, headers=headers)
resp.raise_for_status()
logger.debug("HTTP audit event forwarded to %s (status %s)", url, resp.status_code)
+1
View File
@@ -32,6 +32,7 @@ _TABLE_ORDER = [
"processing_logs",
"application_settings",
"settings_audit_log",
"audit_logs",
"saved_searches",
"webhook_configs",
]
+2
View File
@@ -6,6 +6,7 @@ from fastapi import APIRouter
from app.views.admin_users import router as admin_users_router
from app.views.api_tokens import router as api_tokens_router
from app.views.audit_logs import router as audit_logs_router
from app.views.backup import router as backup_router
from app.views.db_wizard import router as db_wizard_router
from app.views.dropbox import router as dropbox_router
@@ -60,4 +61,5 @@ router.include_router(imap_accounts_router) # Per-user IMAP ingestion accounts
router.include_router(integrations_router) # Unified integrations dashboard
router.include_router(notifications_router) # User notification dashboard
router.include_router(scheduled_jobs_router) # Admin scheduled batch jobs
router.include_router(audit_logs_router) # Comprehensive audit log viewer
router.include_router(help_router) # Built-in help / How-To docs
+46
View File
@@ -0,0 +1,46 @@
"""
Audit log viewer UI — admin-only page with filtering and SIEM status.
"""
import logging
from fastapi import Depends, HTTPException, Request, status
from sqlalchemy.orm import Session
from app.views.base import APIRouter, get_db, require_login, settings, templates
from app.views.settings import require_admin_access
logger = logging.getLogger(__name__)
router = APIRouter()
@router.get("/admin/audit-logs")
@require_login
@require_admin_access
async def audit_logs_page(request: Request, db: Session = Depends(get_db)):
"""Comprehensive audit log viewer with filtering controls.
Displays a chronological log of all significant actions: logins,
document operations, settings changes, and admin actions. The
actual data is fetched client-side via the ``/api/audit-logs`` JSON
endpoint so that filters, pagination, and live refresh work without
full-page reloads.
"""
try:
siem_enabled = settings.audit_siem_enabled
siem_transport = settings.audit_siem_transport if siem_enabled else None
return templates.TemplateResponse(
"audit_logs.html",
{
"request": request,
"app_version": settings.version,
"siem_enabled": siem_enabled,
"siem_transport": siem_transport,
},
)
except Exception as e:
logger.error("Error loading audit logs page: %s", e)
raise HTTPException(
status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
detail="Failed to load audit logs page",
)