homelab_automation/app/routes/metrics.py

"""
Routes API pour les métriques système et des hôtes.
"""

from datetime import timezone
from typing import Optional

from fastapi import APIRouter, Depends, HTTPException
from pydantic import BaseModel
from sqlalchemy.ext.asyncio import AsyncSession
import pytz

from app.core.dependencies import get_db, verify_api_key
from app.core.config import settings


from app.crud.host_metrics import HostMetricsRepository
from app.crud.app_setting import AppSettingRepository
from app.services import db


class CollectionScheduleRequest(BaseModel):
    interval: str = "off"

router = APIRouter()


def _to_app_timezone(dt):
    if dt is None:
        return None
    if getattr(dt, "tzinfo", None) is None:
        dt = dt.replace(tzinfo=timezone.utc)
    tz = pytz.timezone(settings.scheduler_timezone)
    return dt.astimezone(tz)


@router.get("")
async def get_metrics(api_key_valid: bool = Depends(verify_api_key)):
    """Récupère les métriques système globales."""
    return db.metrics


@router.get("/all-hosts")
async def get_all_hosts_metrics(
    api_key_valid: bool = Depends(verify_api_key),
    db_session: AsyncSession = Depends(get_db)
):
    """Récupère les métriques de tous les hôtes."""
    try:
        repo = HostMetricsRepository(db_session)
        metrics_dict = await repo.get_all_latest()
        
        # Convertir en dict host_id -> metrics
        result = {}
        for host_id, m in metrics_dict.items():
            collected_at = _to_app_timezone(m.collected_at)
            result[m.host_id] = {
                "host_id": m.host_id,
                "metric_type": m.metric_type,
                "cpu_usage_percent": m.cpu_usage_percent,
                "cpu_load_1m": m.cpu_load_1m,
                "cpu_model": m.cpu_model,
                "cpu_count": m.cpu_count,
                "cpu_cores": getattr(m, "cpu_cores", None),
                "cpu_threads": getattr(m, "cpu_threads", None),
                "cpu_max_mhz": getattr(m, "cpu_max_mhz", None),
                "cpu_temperature": getattr(m, "cpu_temperature", None),
                "memory_usage_percent": m.memory_usage_percent,
                "memory_total_mb": m.memory_total_mb,
                "memory_used_mb": m.memory_used_mb,
                "disk_root_usage_percent": m.disk_root_usage_percent,
                "disk_root_total_gb": m.disk_root_total_gb,
                "disk_root_used_gb": m.disk_root_used_gb,
                "disk_info": getattr(m, "disk_info", None),
                "disk_devices": getattr(m, "disk_devices", None),
                "lvm_info": getattr(m, "lvm_info", None),
                "zfs_info": getattr(m, "zfs_info", None),
                "storage_details": getattr(m, "storage_details", None),
                "os_name": m.os_name,
                "uptime_human": m.uptime_human,
                "last_collected": collected_at,
                "collected_at": collected_at,
                "collection_status": "success" if not m.error_message else "failed",
                "error_message": m.error_message,
            }
        return result
    except Exception as e:
        # Si la table n'existe pas encore ou autre erreur
        return {}


@router.get("/collection-schedule")
async def get_collection_schedule(
    api_key_valid: bool = Depends(verify_api_key),
    db_session: AsyncSession = Depends(get_db)
):
    """Récupère l'intervalle de collecte des métriques."""
    try:
        repo = AppSettingRepository(db_session)
        setting = await repo.get("metrics_collection_interval")
        interval = setting.value if setting else "off"
        return {"interval": interval}
    except Exception:
        return {"interval": "off"}


@router.post("/collection-schedule")
async def set_collection_schedule(
    request: CollectionScheduleRequest,
    api_key_valid: bool = Depends(verify_api_key),
    db_session: AsyncSession = Depends(get_db)
):
    """Définit l'intervalle de collecte des métriques."""
    interval = request.interval
    valid_intervals = ["off", "5min", "15min", "30min", "1h", "6h", "12h", "24h"]
    if interval not in valid_intervals:
        raise HTTPException(
            status_code=400,
            detail=f"Intervalle invalide. Valeurs acceptées: {valid_intervals}"
        )
    
    try:
        repo = AppSettingRepository(db_session)
        await repo.set("metrics_collection_interval", interval)
        await db_session.commit()
        return {"interval": interval, "message": f"Intervalle de collecte défini à {interval}"}
    except Exception as e:
        raise HTTPException(status_code=500, detail=str(e))


@router.get("/{host_id}")
async def get_host_metrics(
    host_id: str,
    api_key_valid: bool = Depends(verify_api_key),
    db_session: AsyncSession = Depends(get_db)
):
    """Récupère les métriques d'un hôte spécifique."""
    try:
        repo = HostMetricsRepository(db_session)
        metrics = await repo.get_latest_for_host(host_id)
        
        if not metrics:
            return {"host_id": host_id, "collection_status": "no_data"}

        collected_at = _to_app_timezone(metrics.collected_at)
        
        return {
            "host_id": metrics.host_id,
            "metric_type": metrics.metric_type,
            "cpu_usage_percent": metrics.cpu_usage_percent,
            "cpu_load_1m": metrics.cpu_load_1m,
            "cpu_model": metrics.cpu_model,
            "cpu_count": metrics.cpu_count,
            "memory_usage_percent": metrics.memory_usage_percent,
            "memory_total_mb": metrics.memory_total_mb,
            "memory_used_mb": metrics.memory_used_mb,
            "disk_root_usage_percent": metrics.disk_root_usage_percent,
            "disk_root_total_gb": metrics.disk_root_total_gb,
            "disk_root_used_gb": metrics.disk_root_used_gb,
            "disk_info": metrics.disk_info,
            "os_name": metrics.os_name,
            "uptime_human": metrics.uptime_human,
            "last_collected": collected_at,
            "collected_at": collected_at,
            "collection_status": "success" if not metrics.error_message else "failed",
            "error_message": metrics.error_message,
        }
    except Exception as e:
        return {"host_id": host_id, "collection_status": "error", "error_message": str(e)}