Add service health checks and dynamic Apps page

Backend polls each registered service's /health endpoint every 30 s via a
background asyncio task. GET /api/services exposes the live status snapshot.
The Apps page now renders from this endpoint — showing "Unavailable" (dimmed,
non-clickable) when a service is registered but its container is unreachable.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
curo1305
2026-04-17 17:31:36 +02:00
parent 1f8f866414
commit 3248607790
10 changed files with 278 additions and 80 deletions
+13 -5
View File
@@ -40,6 +40,14 @@ JWT signing uses a 4096-bit RSA key pair (`RS256`). Keys are generated by `scrip
| `GET` | `/api/admin/users` | List all users (admin only) |
| `PATCH` | `/api/admin/users/{id}` | Update user (role, active flag) |
### Services (`/api/services`)
| Method | Path | Description |
|--------|------|-------------|
| `GET` | `/api/services` | Returns health status of all registered feature services |
A background task (`service_health.py`) polls each service's `/health` endpoint every 30 s and stores the result in memory. The first check runs immediately on startup. Any authenticated user may call `GET /api/services`; the frontend uses it to drive app card visibility.
### Settings (`/api/settings`)
| Method | Path | Description |
@@ -80,11 +88,11 @@ Browser (port 5173 dev / 80 prod)
└── /api/* → backend:8000 (FastAPI)
┌───────────┼────────────┐
/auth /settings /documents/*
/users (JSON │
/admin volume) └── proxy → doc-service:8001
/profile
┌───────────┼────────────┬──────────────
/auth /settings /documents/* /services
/users (JSON │
/admin volume) └── proxy → health-check loop
/profile doc-service:8001 (30s poll)
```
---
+1
View File
@@ -15,6 +15,7 @@ class Settings(BaseSettings):
CORS_ORIGINS: list[str] = ["http://localhost:5173"]
DOC_SERVICE_URL: str = "http://doc-service:8001"
AI_SERVICE_URL: str = "http://ai-service:8010"
@field_validator("JWT_PRIVATE_KEY", "JWT_PUBLIC_KEY", mode="before")
+25 -2
View File
@@ -1,11 +1,33 @@
import asyncio
from contextlib import asynccontextmanager
from fastapi import FastAPI
from fastapi.middleware.cors import CORSMiddleware
from app.core.config import settings
from app.routers import admin, auth, categories_proxy, documents_proxy, profile, users
from app.routers import admin, auth, categories_proxy, documents_proxy, profile, services, users
from app.routers import settings as settings_router
from app.services.service_health import check_all, health_check_loop, register_services
app = FastAPI(title=settings.PROJECT_NAME, version="0.1.0")
@asynccontextmanager
async def lifespan(app: FastAPI):
register_services(
doc_service_url=settings.DOC_SERVICE_URL,
ai_service_url=settings.AI_SERVICE_URL,
)
# Run an initial check immediately so the first API response is accurate
await check_all()
task = asyncio.create_task(health_check_loop())
yield
task.cancel()
try:
await task
except asyncio.CancelledError:
pass
app = FastAPI(title=settings.PROJECT_NAME, version="0.1.0", lifespan=lifespan)
app.add_middleware(
CORSMiddleware,
@@ -20,6 +42,7 @@ app.include_router(users.router, prefix="/api/users", tags=["users"])
app.include_router(profile.router, prefix="/api/profile", tags=["profile"])
app.include_router(admin.router, prefix="/api/admin", tags=["admin"])
app.include_router(settings_router.router, prefix="/api/settings", tags=["settings"])
app.include_router(services.router, prefix="/api/services", tags=["services"])
# categories_proxy MUST be registered before documents_proxy —
# otherwise /api/documents/{path:path} swallows /api/documents/categories/*
app.include_router(
+22
View File
@@ -0,0 +1,22 @@
"""
GET /api/services — returns health status for all registered feature services.
Available to any authenticated user so the frontend can drive app visibility.
"""
from fastapi import APIRouter, Depends
from app.deps import get_current_user
from app.models.user import User
from app.services.service_health import get_all_statuses
router = APIRouter()
@router.get("")
async def list_services(_: User = Depends(get_current_user)) -> list[dict]:
"""
Returns each registered service with its current health status.
healthy=true → service responded 200 on its last /health poll
healthy=false → unreachable, timed out, or not yet polled
"""
return get_all_statuses()
View File
+108
View File
@@ -0,0 +1,108 @@
"""
Background health-checker for registered feature services.
Polls each service's /health endpoint every POLL_INTERVAL seconds and stores
the result in an in-memory dict. The REST layer reads from that dict — no DB,
no blocking calls on the request path.
"""
import asyncio
import logging
from dataclasses import dataclass, field
import httpx
logger = logging.getLogger(__name__)
POLL_INTERVAL = 30 # seconds
@dataclass
class ServiceDefinition:
id: str
name: str
description: str
internal_url: str # e.g. http://doc-service:8001
health_path: str = "/health"
app_path: str = "" # frontend route; empty = no open button
settings_path: str = "" # frontend admin-settings route
# ── Registry ──────────────────────────────────────────────────────────────────
# Add new services here. The internal_url is filled in at startup from settings.
_REGISTRY: list[ServiceDefinition] = []
# id → True/False/None (None = not yet checked)
_health: dict[str, bool | None] = {}
def register_services(doc_service_url: str, ai_service_url: str) -> None:
"""Called once during app startup to populate the registry from config."""
global _REGISTRY, _health
_REGISTRY = [
ServiceDefinition(
id="doc-service",
name="Documents",
description="Upload PDF files, extract data, and organise them with categories.",
internal_url=doc_service_url,
health_path="/health",
app_path="/apps/documents",
settings_path="/apps/documents/settings/admin",
),
ServiceDefinition(
id="ai-service",
name="AI Service",
description="Shared AI provider for all features. Configure model, credentials, and connection.",
internal_url=ai_service_url,
health_path="/health",
app_path="",
settings_path="/apps/ai/settings/admin",
),
]
_health = {svc.id: None for svc in _REGISTRY}
# ── Health check logic ────────────────────────────────────────────────────────
async def _check_service(svc: ServiceDefinition) -> None:
url = f"{svc.internal_url}{svc.health_path}"
try:
async with httpx.AsyncClient(timeout=5.0) as client:
resp = await client.get(url)
_health[svc.id] = resp.status_code == 200
except Exception:
_health[svc.id] = False
async def check_all() -> None:
"""Run health checks for all registered services concurrently."""
await asyncio.gather(*[_check_service(svc) for svc in _REGISTRY])
async def health_check_loop() -> None:
"""Runs forever; polls every POLL_INTERVAL seconds."""
while True:
await check_all()
await asyncio.sleep(POLL_INTERVAL)
# ── Public read API ───────────────────────────────────────────────────────────
def get_all_statuses() -> list[dict]:
"""Return the current health snapshot for all registered services."""
return [
{
"id": svc.id,
"name": svc.name,
"description": svc.description,
"app_path": svc.app_path,
"settings_path": svc.settings_path,
# None means not yet checked; treat as unhealthy for the UI
"healthy": bool(_health.get(svc.id)),
}
for svc in _REGISTRY
]
@@ -0,0 +1,22 @@
# 2026-04-17 — Service health checks and dynamic Apps page
**Timestamp:** 2026-04-17T00:00:00Z
## Summary
Added a background health-check system to the backend that polls each registered feature service every 30 seconds. The Apps page now renders dynamically based on live service status — showing "Unavailable" when a service is registered but its container is unreachable.
## Files Added / Modified / Deleted
### Added
- `backend/app/services/__init__.py` — package init
- `backend/app/services/service_health.py` — service registry, background polling loop (`POLL_INTERVAL=30s`), `get_all_statuses()` read API
- `backend/app/routers/services.py``GET /api/services` endpoint (requires auth)
### Modified
- `backend/app/core/config.py` — added `DOC_SERVICE_URL` setting (default `http://doc-service:8001`)
- `backend/app/main.py` — added FastAPI `lifespan` context manager: registers services, runs initial health check, starts background polling loop; mounts `/api/services` router
- `frontend/src/api/client.ts` — added `ServiceStatus` interface and `getServices()` API function
- `frontend/src/pages/AppsPage.tsx` — replaced hardcoded `APPS` array with dynamic query to `GET /api/services`; adds "Unavailable" state with dimmed card and explanation text
- `backend/STATUS.md` — documented `/api/services` endpoint and health-check architecture
- `frontend/STATUS.md` — documented dynamic Apps page behaviour
+5 -3
View File
@@ -35,9 +35,11 @@ All API calls go through `src/api/client.ts` (single Axios instance, JWT injecte
### Apps page (`/apps`)
Cards for each installed app:
- **Documents** — link to `/apps/documents`; admin gear icon → `/apps/documents/settings/admin`
- **AI Service** — infrastructure card; admin gear icon → `/apps/ai/settings/admin`; no Open button (no user-facing UI)
Cards are rendered dynamically from `GET /api/services` (polled every 30 s via TanStack Query):
- **healthy=true + app_path set** — clickable card with "Available" badge
- **healthy=true + no app_path** — non-clickable card (e.g. AI Service — no user UI)
- **healthy=false** — non-clickable, dimmed card with "Unavailable" badge and explanation text
- Admin settings link shown for admins regardless of health status
### Sidebar navigation
+13
View File
@@ -224,6 +224,19 @@ export const updateDocumentLimits = (max_pdf_mb: number) =>
export const getDocumentLimits = () =>
api.get<Record<string, unknown>>("/settings/documents/limits").then((r) => r.data);
// --- Services ---
export interface ServiceStatus {
id: string;
name: string;
description: string;
healthy: boolean;
app_path: string;
settings_path: string;
}
export const getServices = () =>
api.get<ServiceStatus[]>("/services").then((r) => r.data);
// --- System Prompts (admin only) ---
export interface ServiceSystemPrompt {
label: string;
+69 -70
View File
@@ -1,84 +1,84 @@
import { Link } from "react-router-dom";
import { useQuery } from "@tanstack/react-query";
import { getMe } from "../api/client";
import { getMe, getServices } from "../api/client";
interface AppCard {
slug: string;
name: string;
description: string;
status: "available" | "coming_soon";
path: string;
settingsPath?: string;
}
const cardBase: React.CSSProperties = {
backgroundColor: "rgb(var(--color-surface))",
border: "1px solid rgb(var(--color-border))",
borderRadius: 8,
padding: 24,
width: 280,
display: "flex",
flexDirection: "column",
gap: 12,
};
const APPS: AppCard[] = [
{
slug: "documents",
name: "Documents",
description: "Upload PDF files, extract data, and organise them with categories.",
status: "available",
path: "/apps/documents",
settingsPath: "/apps/documents/settings/admin",
},
{
slug: "ai",
name: "AI Service",
description: "Shared AI provider for all features. Configure model, credentials, and connection.",
status: "available",
path: "",
settingsPath: "/apps/ai/settings/admin",
},
];
const clickableCard: React.CSSProperties = {
...cardBase,
cursor: "pointer",
textDecoration: "none",
color: "inherit",
transition: "box-shadow 150ms, border-color 150ms",
};
const unavailableCard: React.CSSProperties = {
...cardBase,
opacity: 0.6,
};
export default function AppsPage() {
const { data: user } = useQuery({ queryKey: ["me"], queryFn: getMe });
const { data: services = [] } = useQuery({
queryKey: ["services"],
queryFn: getServices,
refetchInterval: 30_000,
});
return (
<>
<div style={{ padding: 32, maxWidth: 900, margin: "0 auto" }}>
<h1>Apps</h1>
<div style={{ display: "flex", gap: 24, flexWrap: "wrap", marginTop: 24 }}>
{APPS.map((app) => (
<div
key={app.slug}
style={{
border: "1px solid #ddd",
borderRadius: 8,
padding: 24,
width: 280,
display: "flex",
flexDirection: "column",
gap: 12,
}}
>
<div style={{ padding: 32, maxWidth: 900, margin: "0 auto" }}>
<h1>Apps</h1>
<div style={{ display: "flex", gap: 24, flexWrap: "wrap", marginTop: 24 }}>
{services.map((svc) => {
const canOpen = svc.healthy && !!svc.app_path;
const CardWrapper = canOpen ? Link : "div";
const wrapperProps = canOpen
? {
to: svc.app_path,
style: clickableCard,
onMouseEnter: (e: React.MouseEvent<HTMLAnchorElement>) => {
e.currentTarget.style.boxShadow = "0 4px 12px rgb(0 0 0 / 0.12)";
e.currentTarget.style.borderColor = "rgb(var(--color-primary))";
},
onMouseLeave: (e: React.MouseEvent<HTMLAnchorElement>) => {
e.currentTarget.style.boxShadow = "";
e.currentTarget.style.borderColor = "rgb(var(--color-border))";
},
}
: { style: svc.healthy ? cardBase : unavailableCard };
return (
<CardWrapper key={svc.id} {...(wrapperProps as any)}>
<div style={{ display: "flex", justifyContent: "space-between", alignItems: "center" }}>
<h2 style={{ margin: 0, fontSize: 18 }}>{app.name}</h2>
{app.status === "available" ? (
<h2 style={{ margin: 0, fontSize: 18 }}>{svc.name}</h2>
{svc.healthy ? (
<span style={{ fontSize: 12, color: "#2a9d8f", fontWeight: 600 }}>Available</span>
) : (
<span style={{ fontSize: 12, color: "#aaa" }}>Coming soon</span>
<span style={{ fontSize: 12, color: "#e76f51", fontWeight: 600 }}>Unavailable</span>
)}
</div>
<p style={{ margin: 0, color: "#555", fontSize: 14 }}>{app.description}</p>
<p style={{ margin: 0, color: "rgb(var(--color-text-muted))", fontSize: 14 }}>
{svc.description}
</p>
{!svc.healthy && (
<p style={{ margin: 0, fontSize: 12, color: "#e76f51" }}>
This service is currently unavailable. Please try again later or contact your administrator.
</p>
)}
<div style={{ display: "flex", gap: 8, marginTop: "auto" }}>
{app.status === "available" && app.path && (
{user?.is_admin && svc.settings_path && (
<Link
to={app.path}
style={{
padding: "6px 14px",
background: "#222",
color: "#fff",
borderRadius: 4,
textDecoration: "none",
fontSize: 14,
}}
>
Open
</Link>
)}
{user?.is_admin && app.settingsPath && app.status === "available" && (
<Link
to={app.settingsPath}
to={svc.settings_path}
onClick={(e) => e.stopPropagation()}
style={{
padding: "6px 14px",
border: "1px solid #ccc",
@@ -93,11 +93,10 @@ export default function AppsPage() {
</Link>
)}
</div>
</div>
))}
</div>
</CardWrapper>
);
})}
</div>
</>
</div>
);
}