Self-hosted monitoring for homelabs: host and container metrics and uptime checks from all your machines in one dashboard, with an optional AI agent that diagnoses incidents and suggests fixes.
Python ai-agentsaiopsdashboarddockerfastapihomelabincident-managementllmmonitoringopenaireactroot-cause-analysisself-hosteduptime-monitoring
import os
from datetime import datetime, timedelta, timezone
os.environ.setdefault("DATABASE_URL", "sqlite://")
os.environ.setdefault("REDIS_URL", "redis://localhost:6379/0")
import pytest
from fastapi import HTTPException
from sqlalchemy import create_engine
from sqlalchemy.orm import sessionmaker
from src.collector import collector
from src.backend.routes import monitoring as monitoring_routes
from src.backend import main
from src.backend.models import Base, HostMetric, Server, User
@pytest.fixture
def db(monkeypatch):
engine = create_engine("sqlite://")
Base.metadata.create_all(engine)
session = sessionmaker(bind=engine)()
session.add(User(id=1, name="Owner", email="[email protected]", password_hash="x", role="admin"))
now = datetime.now(timezone.utc)
for server_id, name, age in ((1, "local", 3), (2, "nas", 2), (3, "pi", 1)):
session.add(Server(id=server_id, owner_id=1, hostname=name, check_interval_seconds=60, created_at=now - timedelta(days=age), collector_token_hash=main._hash_secret(f"token-{server_id}")))
session.commit()
monkeypatch.setattr(main, "_current_user", lambda session_id, db: {"id": 1})
return session
def _ingest(db, server_id, **server):
monitoring_routes.ingest_metrics(main.MetricsIngestRequest(server_id=server_id, server=server, host={"cpu_utilization_pct": 10, "ram_pct": 20, "disk_usage_pct": 30}), authorization=f"Bearer token-{server_id}", db=db)
def test_dashboard_host_is_listed_first_and_flagged(db):
_ingest(db, 1, dashboard_host=True)
_ingest(db, 2)
servers = monitoring_routes.monitoring_overview(session_id="s", db=db)["servers"]
assert [(s["hostname"], s["is_dashboard_host"]) for s in servers] == [("local", True), ("pi", False), ("nas", False)]
def test_only_latest_claimant_is_dashboard_host_and_flag_can_be_dropped(db):
_ingest(db, 1, dashboard_host=True)
_ingest(db, 2, dashboard_host=True)
assert [s.id for s in db.query(Server).filter_by(is_dashboard_host=True)] == [2]
_ingest(db, 2)
assert db.query(Server).filter_by(is_dashboard_host=True).count() == 0
def test_history_returns_recent_samples_oldest_first(db):
now = datetime.now(timezone.utc)
for minutes_ago, cpu in ((90, 1), (20, 2), (5, 3)):
db.add(HostMetric(server_id=1, time=now - timedelta(minutes=minutes_ago), cpu_utilization_pct=cpu, ram_pct=50, disk_usage_pct=60))
db.add(HostMetric(server_id=2, time=now, cpu_utilization_pct=99))
db.commit()
samples = monitoring_routes.server_history(1, minutes=30, session_id="s", db=db)["samples"]
assert [(s["cpu"], s["ram"], s["disk"]) for s in samples] == [(2, 50, 60), (3, 50, 60)]
def test_history_of_unknown_server_is_404(db):
with pytest.raises(HTTPException) as exc:
monitoring_routes.server_history(99, session_id="s", db=db)
assert exc.value.status_code == 404
def test_collector_sends_flag_only_when_configured(monkeypatch):
monkeypatch.setenv("INFRADECK_DASHBOARD_HOST", "true")
assert collector.collect_host()[0]["dashboard_host"] is True
monkeypatch.delenv("INFRADECK_DASHBOARD_HOST")
assert "dashboard_host" not in collector.collect_host()[0]
def _ingest_containers(db, server_id, names):
containers = [{"container_id": f"id-{name}", "name": name, "status": "running"} for name in names]
monitoring_routes.ingest_metrics(main.MetricsIngestRequest(server_id=server_id, host={"cpu_utilization_pct": 1}, containers=containers), authorization=f"Bearer token-{server_id}", db=db)
def _overview_server(db, server_id):
return next(s for s in monitoring_routes.monitoring_overview(session_id="s", db=db)["servers"] if s["id"] == server_id)
def test_overview_lists_only_containers_from_latest_check_in(db):
_ingest_containers(db, 2, ["web", "old"])
_ingest_containers(db, 2, ["web", "db"])
assert sorted(c["name"] for c in _overview_server(db, 2)["containers"]) == ["db", "web"]
def test_overview_lists_no_containers_when_latest_check_in_had_none(db):
_ingest_containers(db, 2, ["web"])
db.query(main.ContainerMetric).update({"time": datetime.now(timezone.utc) - timedelta(minutes=5)})
db.commit()
_ingest_containers(db, 2, [])
assert _overview_server(db, 2)["containers"] == []
def test_overview_counts_active_alerts_per_server(db):
from src.backend import alerts
alerts.raise_alert(db, 2, "a", alerts.WARNING, "t", "d")
alerts.raise_alert(db, 2, "b", alerts.WARNING, "t", "d")
alerts.raise_alert(db, 3, "c", alerts.WARNING, "t", "d")
alerts.resolve_alerts(db, 3, ["c"])
db.commit()
assert _overview_server(db, 2)["active_alert_count"] == 2
assert _overview_server(db, 3)["active_alert_count"] == 0