Files
manage/backend/src/media_library_viewer_api/routers/monitoring.py
T
2026-07-14 20:58:46 +00:00

231 lines
8.2 KiB
Python

"""Monitoring router — observability service status.
Observability components (Alertmanager, Prometheus) are resolved from
the service registry, not environment variables. The endpoints pick the first
enabled instance of a type when no ``service_id`` is given, and return graceful
"not configured" / "unreachable" payloads so the UI always renders a health card.
"""
from __future__ import annotations
import logging
from typing import Any
import requests
from fastapi import APIRouter, Body, Depends
from media_library_viewer_api.clients.http_timeout import http_timeout
from media_library_viewer_api.dependencies import get_settings_store
from media_library_viewer_api.services.service_resolution import resolve_service_record
from media_library_viewer_api.services.settings_store import SettingsStore
from media_library_viewer_api.widgets.sources import ServiceRecord
logger = logging.getLogger(__name__)
def _base_url(service: ServiceRecord) -> str:
return str(service.config.get("base_url") or "").rstrip("/")
def _timeout(service: ServiceRecord, default: int) -> tuple[float, float]:
"""Return a (connect, read) timeout tuple from the service config."""
read = int(service.config.get("timeout_seconds") or default)
return http_timeout(read)
def _auth_headers(service: ServiceRecord) -> dict[str, str]:
api_key = str(service.secrets.get("api_key") or "")
return {"Authorization": f"Bearer {api_key}"} if api_key else {}
def _status_response(service: ServiceRecord | None, *, version: str = "", error: str | None = None) -> dict[str, Any]:
return {
"up": error is None,
"version": version or "",
"service_id": service.id if service else "",
"name": service.name if service else "",
"error": error,
}
def _summary_from_alerts(alerts: list[dict[str, Any]]) -> dict[str, Any]:
"""Build a UI-friendly summary from Alertmanager /api/v1/alerts payload."""
from media_library_viewer_api.integrations.alertmanager import summarize_alerts
return summarize_alerts(alerts)
router = APIRouter(prefix="/api/monitoring", tags=["monitoring"])
@router.get("/alerts")
def get_alertmanager_alerts(
service_id: str | None = None,
store: SettingsStore = Depends(get_settings_store),
) -> dict[str, Any]:
"""Return a summary of active Alertmanager alerts for the UI.
Resolves an ``alertmanager`` service instance from the registry. When none
is configured the endpoint returns an empty summary with an
``alertmanager_not_configured`` error so the UI can render a health card.
"""
service = resolve_service_record(store, "alertmanager", service_id)
if service is None:
return {"total": 0, "by_severity": {}, "alerts": [], "error": "alertmanager_not_configured"}
try:
response = requests.get(
f"{_base_url(service)}/api/v1/alerts",
headers=_auth_headers(service),
timeout=_timeout(service, 5),
)
response.raise_for_status()
data = response.json()
except Exception:
logger.exception("Failed to fetch Alertmanager alerts")
return {
"total": 0,
"by_severity": {},
"alerts": [],
"error": "alertmanager_unreachable",
"service_id": service.id,
"name": service.name,
}
if data.get("status") != "success":
return {
"total": 0,
"by_severity": {},
"alerts": [],
"error": data.get("error", "unknown"),
"service_id": service.id,
"name": service.name,
}
summary = _summary_from_alerts(data.get("data", []))
summary["service_id"] = service.id
summary["name"] = service.name
logger.info("Alertmanager alerts requested total=%s", summary["total"])
return summary
@router.get("/alertmanager-status")
def get_alertmanager_status(
service_id: str | None = None,
store: SettingsStore = Depends(get_settings_store),
) -> dict[str, Any]:
"""Return Alertmanager cluster/status for the UI health card."""
service = resolve_service_record(store, "alertmanager", service_id)
if service is None:
return {
"up": False,
"version": "",
"uptime": "",
"name": "",
"peers": [],
"error": "alertmanager_not_configured",
}
try:
response = requests.get(
f"{_base_url(service)}/api/v2/status",
headers=_auth_headers(service),
timeout=_timeout(service, 5),
)
response.raise_for_status()
data = response.json()
except Exception:
logger.exception("Failed to fetch Alertmanager status")
return {
"up": False,
"version": "",
"uptime": "",
"name": service.name,
"peers": [],
"service_id": service.id,
"error": "alertmanager_unreachable",
}
cluster = data.get("cluster") or {}
status = data.get("clusterStatus") or {}
return {
"up": True,
"version": data.get("versionInfo", {}).get("version", ""),
"uptime": status.get("createdAt", ""),
"name": service.name,
"peers": [p.get("name", "") for p in cluster.get("peers", [])],
"service_id": service.id,
"error": None,
}
@router.get("/prometheus-status")
def get_prometheus_status(
service_id: str | None = None,
store: SettingsStore = Depends(get_settings_store),
) -> dict[str, Any]:
"""Probe a Prometheus service's health via the Grafana gateway path (GM-110).
Issues a trivial ``up`` query through Grafana ``/api/ds/query``. Success
validates the full path: Grafana is reachable, the API key works, and the
Prometheus datasource responds.
"""
service = resolve_service_record(store, "prometheus", service_id)
if service is None:
return _status_response(None, error="no_service_configured")
grafana_url = str(service.config.get("grafana_url") or "").rstrip("/")
api_key = str(service.secrets.get("grafana_api_key") or "")
datasource_uid = str(service.config.get("datasource_uid") or "prometheus")
timeout = int(service.config.get("timeout_seconds") or 60)
if not grafana_url or not api_key:
return _status_response(service, error="gateway_not_configured")
body = {
"queries": [
{
"datasource": {"uid": datasource_uid, "type": "prometheus"},
"expr": "up",
"format": "time_series",
"intervalMs": 15_000,
"maxDataPoints": 1,
"refId": "A",
}
],
"from": "now-1m",
"to": "now",
}
try:
resp = requests.post(
f"{grafana_url}/api/ds/query",
json=body,
headers={"Authorization": f"Bearer {api_key}", "Content-Type": "application/json"},
timeout=http_timeout(timeout),
)
resp.raise_for_status()
except requests.HTTPError as exc:
status_code = exc.response.status_code if exc.response else 0
if status_code in (401, 403):
return _status_response(service, error="auth_failed")
return _status_response(service, error="gateway_error")
except requests.RequestException:
logger.exception("Failed to fetch Prometheus status via gateway")
return _status_response(service, error="prometheus_unreachable")
return _status_response(service, version="ok")
@router.post("/alertmanager-webhook")
def receive_alertmanager_webhook(payload: dict[str, Any] = Body(...)) -> dict[str, str]:
"""Receive alerts from Alertmanager and log them for audit/debug.
This endpoint is the receiver referenced by the optional ``webhook_configs``
block in Alertmanager. It is log-only: received payloads are recorded but not
forwarded anywhere. (The previous outbound relay to ``ALERTMANAGER_WEBHOOK_URL``
was removed when observability became service-registry configured.)
"""
alerts = payload.get("alerts", [])
logger.info(
"Received Alertmanager webhook with %s alert(s), status=%s",
len(alerts),
payload.get("status", "unknown"),
)
return {"status": "received"}