Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/sheeplettuce/Monitor/llms.txt

Use this file to discover all available pages before exploring further.

Monitor API has a built-in observability layer that runs automatically and is also queryable on demand. On startup, the server executes a suite of health checks and logs the results; those same checks repeat every 5 minutes throughout the process lifetime. A dedicated REST endpoint exposes a real-time snapshot of the same data so that external tools, dashboards, and client apps can poll the service programmatically.

​
Health endpoint

GET /api/health No authentication is required. Returns a JSON object describing the current state of every monitored subsystem. 
curl http://localhost:3000/api/health
Example response: 
{
  "ip": "192.168.1.42",
  "backend": true,
  "database": true,
  "disco": {
    "libre_GB": "45.2",
    "usado_GB": "14.8",
    "total_GB": "60.0"
  },
  "evidencias": {
    "archivos": 312,
    "tamaño_MB": "1240.5"
  },
  "expedientes": 87,
  "ram": {
    "usada_MB": 128,
    "total_MB": 8192
  }
}
Field reference:

ip

The IP address of the client that made the request, as reported by Express (req.ip). This reflects the caller’s address, not the server’s address.

backend

Always true when this endpoint responds. Confirms the Express process is running and reachable on the network.

database

true if PostgreSQL responds successfully to a SELECT 1 query; false if the connection times out or throws an error. A false value here means the API is running but cannot read or write data.

disco

Disk space (in GB, as strings) for the drive where the server process is running. Reported as libre_GB, usado_GB, and total_GB. On Windows the C: drive is queried via wmic; on Linux/macOS df -BGB . reports the filesystem hosting the current working directory. If free space drops below 10 GB, a WARN entry is written to the log.

evidencias

archivos: count of top-level entries in EVIDENCIAS_DIR. tamaño_MB: combined size of those entries in megabytes (as a string with one decimal place). Reflects the growth of uploaded evidence files over time.

expedientes

Total count of expediente rows in the database. Gives a quick sense of overall system load without running a separate query.

ram

usada_MB: RSS (resident set size) memory consumed by the Node.js process in megabytes. total_MB: total physical RAM installed on the host machine. Both are rounded integers.
Set up an external uptime monitoring tool such as UptimeRobot or Uptime Kuma to poll GET /api/health on a regular interval (e.g. every 60 seconds). Configure an alert to fire whenever database is false — this is the most critical failure mode because it means the API is accepting requests but cannot persist any data.

​
Service discovery endpoint

GET /discovery Returns the server’s local IPv4 addresses. No authentication required. Clients on the same LAN can call this endpoint to locate the Monitor API server without relying on a hardcoded IP. 
curl http://monitor-server.local/discovery
Response: 
{
  "nombre": "Monitor API",
  "ips": ["192.168.1.10"]
}
This is especially useful during initial setup, when the server’s IP may not yet be known to all clients on the network.

​
mDNS / Bonjour discovery

On startup, Monitor API registers itself with the local network via mDNS (Multicast DNS) using the Bonjour protocol, advertising the service type _monitor._tcp. Clients and operating systems that support mDNS (macOS, most Linux desktops, Windows 10+) can resolve the server by hostname without needing an IP address: 
http://<hostname>.local:3000
This means that once the server is running, any device on the same subnet can reach it using its machine hostname — even if DHCP assigns a different IP on every reboot.

​
Startup checks

When the server process starts, it runs the following checks in sequence and logs the result of each one:
1

checkBackend

Verifies that a process is actively listening on port 3000 (the API port) by inspecting open network connections. On Windows it runs netstat -ano | findstr :3000; on Linux/macOS it runs lsof -i :3000 | grep LISTEN. Logs success if the port is bound, error if nothing is listening.
2

checkDatabase

Executes SELECT 1 against the PostgreSQL database via Prisma. Logs success on a valid response, error with the exception message if the connection fails.
3

checkDisk

Reads available, used, and total disk space for the server’s drive. Logs the figures at info level. If free space is below 10 GB, an additional warn entry is written.
4

checkEvidencias

Reads the EVIDENCIAS_DIR directory and reports the number of entries and their combined size. Logs warn if the directory does not exist (e.g. on a fresh installation before any files have been uploaded).
5

checkFrontend

Makes an HTTP request to http://localhost:8081. Logs success if the frontend responds with 200 OK, warn if it responds with an unexpected status code, and warn (not error) if nothing is listening on that port — this allows the API to run independently without the frontend.
All five checks repeat automatically every 5 minutes while the server is running (setInterval(runStatusChecks, 5 * 60 * 1000)), keeping the log file current without requiring manual intervention.

​
Log file

All log entries — from startup checks, periodic checks, request handlers, and internal services — are appended to logs/app.log in the project root. The file is created automatically if it does not exist. Each line follows this format: 
[ISO timestamp] [LEVEL] [Context] Message | {data}
Example entries: 
[2024-01-13T10:00:00.000Z] [SUCCESS] [PostgreSQL] Conexión exitosa a la base de datos | {}
[2024-01-13T10:00:00.012Z] [INFO] [Disco] Estado del disco local | {"total_GB":"60.0","usado_GB":"14.8","libre_GB":"45.2"}
[2024-01-13T10:00:00.025Z] [WARN] [Frontend] No disponible (sin conexión o no iniciado) | {"puerto":8081}
[2024-01-13T10:00:01.100Z] [SUCCESS] [Expedientes] Expediente creado | {"no_siniestro":"SIN-2024-001"}
Log levels used: SUCCESS, INFO, WARN, ERROR. The log file grows indefinitely — consider setting up log rotation (e.g. with logrotate on Linux) for long-running production deployments.
The EVIDENCIAS_DIR used by the health check is resolved as path.resolve(process.cwd(), process.env["EVIDENCIAS_DIR"] ?? "../evidencias"). If the environment variable is not set, it defaults to ../evidencias relative to process.cwd(). Ensure this path exists before starting the server, or the checkEvidencias step will log a warning on every cycle.

Build docs developers (and LLMs) love

Get started for freeTalk to us