Files
meet/src/backend/core/audit/request.py
T
briquet 4e9f76bec5 ✨(backend) add structured audit logging facility
Add a core.audit package emitting one Elastic Common Schema JSON line per
security-relevant action on a dedicated "audit" logger. Only lasuite.*
and entity.target.raw.* extends on ECS core fields. Each event names
the data stream it belongs to (data_stream.*, event.dataset) and carries
a unique event.id.

The audit logs reads the actor, auth method, tenant and network fields from
the request it is given, else from the request being served, which
AuditLogMiddleware keeps in a context variable so that a service needs
no request argument.

A person is identified by primary key, OIDC sub when the account has one
and email domain, with its privileges as user.roles; an internal peer
calling the backend by service.origin.name. The tenant is the
application client id, else that domain. The client IP is DRF's throttle
ident, which now honours a NUM_PROXIES setting.

The middleware also replaces the inbound X-Request-ID with a fresh id,
unless REQUEST_ID_TRUST_HEADER says the ingress overwrites it, and echoes
it in the response and the gunicorn access log. Audit events carry it as
http.request.id, next to the client's user agent.

An action is defined with a dotted name with its default ECS category
and types, validated at import against the types ECS expects in that
category, so a call site only names what was attempted. A category or
types given to audit.log win over them, and a bare dotted name is
accepted too.

 Meet declares all of the events core/auditing.py:
- audit.register lists the fields describing a model as a target, an
  ECS entity.target, and its ECS entity type. Every target carries its
  model name and primary key, so an unregistered one is still
  identified, and a user target its OIDC sub.
- audit.register_auth_method names the lasuite.auth.method of a DRF
  authentication class

AuditViewMixin audits every response of the CRUD actions a view maps in
audit_actions, and of the extra actions naming theirs with
@action(audit_action=...), from finalize_response: the outcome, reason
and status come from the response. An exception DRF does not handle is recorded
as an internal error before it propagates. The core.audit app also
records Django's login, failed login and logout signals.
2026-10-19 15:16:27 +02:00

115 lines
3.9 KiB
Python

"""Read the network fields and the request id behind an audit event."""
import uuid
from contextvars import ContextVar, Token
from dataclasses import dataclass
from typing import Any
from django.conf import settings
from dockerflow.logging import request_id_context
from rest_framework.throttling import BaseThrottle
USER_AGENT_MAX_LENGTH = 1024
def current_request_id() -> str | None:
"""Return the id of the request being served, if any."""
return request_id_context.get(None)
def resolve_client_ip(request) -> str | None:
"""Return the address of the real client, never the one of a proxy.
It reuses DRF's throttles to identify the client.
"""
return BaseThrottle().get_ident(request) or request.META.get("REMOTE_ADDR")
@dataclass(frozen=True)
class RequestContext:
"""What an audit event reads from the request behind it.
The network fields are read once, when the request comes in. The request
itself is kept for the actor, only known once it is authenticated: it is
the Django request, onto which DRF copies its user and auth, but not its
authenticator, so the auth method of a DRF view is not known from it.
Outside a request, every field is ``None``.
"""
request: Any = None
request_id: str | None = None
method: str | None = None
path: str | None = None
client_ip: str | None = None
user_agent: str | None = None
@classmethod
def from_request(cls, request) -> "RequestContext":
"""Read the context of ``request``, its user agent cut where ECS stops."""
meta = getattr(request, "META", None) or {}
return cls(
request=request,
request_id=current_request_id(),
method=getattr(request, "method", None),
path=getattr(request, "path", None) or None,
client_ip=resolve_client_ip(request),
user_agent=meta.get("HTTP_USER_AGENT", "")[:USER_AGENT_MAX_LENGTH] or None,
)
_request_context: ContextVar[RequestContext | None] = ContextVar(
"audit_request_context", default=None
)
def request_context() -> RequestContext:
"""Return the context of the request being served, as set by ``AuditLogMiddleware``.
It is empty outside a request.
"""
return _request_context.get() or RequestContext()
def set_request_context(context: RequestContext) -> Token:
"""Make ``context`` the one of the request being served, until reset."""
return _request_context.set(context)
def reset_request_context(token: Token) -> None:
"""Restore the context that was current before ``set_request_context``."""
_request_context.reset(token)
class AuditLogMiddleware:
"""Settle the request id and the request context, then echo the id.
It must come right after ``DockerflowMiddleware``, which sets the id from
the inbound ``DOCKERFLOW_REQUEST_ID_HEADER_NAME`` header. Unless
``REQUEST_ID_TRUST_HEADER`` says the ingress overwrites that header, the id
is replaced by a fresh one before anything logs, so that a client, the web
server access log and the audit events of a request can be joined on an id
the client did not choose.
The context of the request is kept while it is served, so that an audit
event emitted far from the view, as from a service, still reads its actor
and network fields from it.
"""
def __init__(self, get_response):
self.get_response = get_response
def __call__(self, request):
if not settings.REQUEST_ID_TRUST_HEADER:
request_id_context.set(str(uuid.uuid4()))
token = set_request_context(RequestContext.from_request(request))
try:
response = self.get_response(request)
finally:
reset_request_context(token)
header = settings.DOCKERFLOW_REQUEST_ID_HEADER_NAME
if not response.has_header(header):
response[header] = current_request_id()
return response