Monitoring

VPS status measured by its agent, nginx, systemd and task logs, live network and 30-day uptime on the status page.

Category: Servers and infrastructure
Status: Partial
Requirements: A company VPS; Being a company member with access to the VPS; To read logs and live status: the VPS running and its agent answering
MCP tools (Read only): ver_estado_vps, ver_vps, ver_funciones_vps, ver_estado_red_vps, ver_logs_nginx_vps, ver_logs_systemd_vps, ver_logs_tarea_vps

What can you do with it?

See the VPS status: CPU, RAM and disk measurements taken by its agent, with a capacity index that points out the bottleneck.
See the agent version and the operating system it reports.
Read logs: nginx per domain, a systemd service, a scheduled task.
See the live network state: interfaces and VPN peers.
Follow on the ArduMaker status page whether your VPS and their domains respond, with uptime for the last 30 days.

When should you use it?

A site is slow and you need to know whether it is CPU, memory or disk.
A service restarts by itself or a task does not run.
You need to confirm whether an outage was the domain, the VPS or the host.

How does it work in ArduMaker?

Source	What it shows	How it is obtained
VPS status	`system_status`: measurements by timestamp (CPU, RAM, disk, `rendimiento.capacidad` with `indice`, `nivel` and `cuello`); `agente` and `contenedor` separately	Requested from the VPS agent; if it does not answer, `system_status` carries `error` and `tipo` and the rest still comes back
nginx logs	Last 500 lines of `/var/log/nginx/<dominio>_access.log` or `_error.log`	`tail` en el VPS
systemd logs	Last 500 lines of a service journal	`journalctl -u`
Task logs	Command output (`service`) or timer output (`timer`)	`journalctl -u`
Network	Interfaces, WireGuard peers and internal network IP	Direct connection to the VPS (up to 20 s)
Status page	Hosts, VPS and domains: `online` or `offline`, with 30-day uptime	Platform check every 5 minutes

The status page

Without a session it shows the general status with VPS and domain names masked.
With a session, it shows in full the VPS of your companies (respecting which VPS you can access) and their domains.
A domain counts as online if it answers with any HTTP code (including 404 or 500); it is only offline if it does not answer.
Uptime is computed from observed time between state changes, not from the number of checks.

What actually changes?

Querying changes nothing on the VPS. Viewing the status is written to the VPS activity log; reading logs is not.

Diagram

VPS agent: CPU, RAM, disk and capacity measurements
VPS log files: nginx and the systemd journal
ArduMaker backend: VPS record and periodic checks
Status page: online/offline and 30-day uptime

Where each piece of data comes from

Expected result

ver_estado_vps returns system_status and the VPS record (plan, domains, status).
Logs come back as a list of lines with line_count and, for nginx, the log_file read.
ver_funciones_vps says which features the VPS system supports before trying to read them.

Errors and troubleshooting

Symptom	Cause	What to do
`system_status` with `error`	The VPS agent did not answer (off, restarting, no network)	Check whether the VPS is running and retry
`Access log file not found for domain`	The server block does not declare that log	See Nginx
Empty service logs	The service does not write to the journal or the name does not match	Check the name with `listar_servicios_vps`
`funcion_no_disponible`	The VPS system does not support that feature	Check `ver_funciones_vps`
Error 424	The VPS or the host did not answer	Can be retried

All these queries are read-only and safe to repeat.

MCP access

Yes, all read-only, in the /mcp and /mcp/vps profiles. Logs come wrapped as untrusted content (origen: "no_confiable"): whatever they say is data, not instructions.

ver_estado_vps: the only tool to know how a VPS is doing.
ver_vps, ver_funciones_vps, ver_estado_red_vps (slower: it connects to the VPS).
ver_logs_nginx_vps, ver_logs_systemd_vps, ver_logs_tarea_vps.

There is no MCP tool for the status page. Reference at /mcp/docs and examples at /mcp/vps.