Healthcheck Monitor

BatAudit pings your service URLs on a schedule and tracks their uptime — closing the loop between "what happened" (audit log) and "is the app still running" (uptime).

When a service goes down or comes back up, BatAudit records the transition as an audit event and fires your configured notifications (Web Push + Webhook).

---

How it works

1. You register a URL and a polling interval per project (e.g. https://api.yourapp.com/health every 60s)
2. The Worker service pings each URL using an HTTP GET with your configured timeout
3. The response status code is compared against the expected status (default: 200)
4. On a UP → DOWN or DOWN → UP transition, BatAudit:
   • Records a system.healthcheck.down or system.healthcheck.up event in the audit log
   • Sends a notification via all active channels (Web Push, Webhook)
5. Results are stored and visible as a history chart per monitor

Notifications fire only on transitions — not on every failed check — so you won't get spammed.

---

Setup

1. Open Settings → Healthcheck in the dashboard
2. Select a project (required — monitors are per-project)
3. Click Add monitor
4. Fill in:
   • Name — e.g. API Health, Payments Service
   • URL — the endpoint to ping (must return HTTP, not HTTPS on localhost in dev)
   • Interval — how often to check, in seconds (minimum: 10s, default: 60s)
   • Timeout — how long to wait for a response (default: 10s)
   • Expected status — the HTTP status code that means "healthy" (default: 200)

---

Dashboard integration

Service breakdown table

The Dashboard service table gains a Health column showing the status of the monitor whose name matches the service_name:

Status	Display
UP	Green dot + "UP"
DOWN	Red dot + "DOWN"
No monitor	—

The monitor Name field should match the service_name sent by your SDK for the column to link them automatically.

Down banner

When any monitor is DOWN, a red alert banner appears at the top of the dashboard listing the affected services and their URLs.

In All Projects view, the Healthcheck settings page shows all DOWN monitors across every project in read-only mode.

---

Monitor actions

Action	Description
Toggle	Pause or resume periodic checks without deleting the monitor
Edit	Update name, URL, interval, timeout, or expected status
Test now	Run an immediate check and see the result inline
History	Expand the row to see the last 50 check results
Delete	Remove the monitor permanently

---

Audit events

Every UP/DOWN transition creates an event in the audit log:

system.healthcheck.down

{
  "event_type": "system.healthcheck.down",
  "service_name": "API Health",
  "path": "https://api.yourapp.com/health",
  "request_body": {
    "url": "https://api.yourapp.com/health",
    "expected_status": 200,
    "status_code": 503,
    "response_ms": 142,
    "error": "expected 200, got 503"
  }
}

system.healthcheck.up

{
  "event_type": "system.healthcheck.up",
  "service_name": "API Health",
  "path": "https://api.yourapp.com/health",
  "request_body": {
    "url": "https://api.yourapp.com/health",
    "status_code": 200,
    "response_ms": 98,
    "downtime_seconds": 312
  }
}

---

API reference

All endpoints require a valid JWT (Authorization: Bearer <token>).

List monitors

GET /v1/monitors?project_id=<id>

Omit project_id to list monitors across all projects.

Create monitor

POST /v1/monitors
Content-Type: application/json

{
  "project_id": "uuid",
  "name": "API Health",
  "url": "https://api.yourapp.com/health",
  "interval_seconds": 60,
  "timeout_seconds": 10,
  "expected_status": 200
}

Maximum 10 monitors per project.

Update monitor

PUT /v1/monitors/:id
Content-Type: application/json

{
  "enabled": false
}

All fields are optional — only provided fields are updated.

Delete monitor

DELETE /v1/monitors/:id

Check history

GET /v1/monitors/:id/history?limit=50

Returns the last N results (max 200).

Test now

POST /v1/monitors/:id/test

Runs an immediate check and returns the result without persisting it.

---

Limits

Limit	Value
Monitors per project	10
Minimum interval	10 seconds
History kept per monitor	200 results
HTTP method used	GET