---
title: "Infrastructure"
description: "The Infrastructures page in Curade Mini tracks each host's health, its CPU, memory and disk use, its metadata and the problems on it and its services."
url: https://curade.ai/en/curade-mini/docs/features/infrastructure/
product: "Curade Mini"
version: "1.0.0"
breadcrumbs:
  - "Curade Mini docs"
  - "Features"
  - "Infrastructure"
---

# Infrastructure

The Infrastructures page in Curade Mini tracks each host's health, its CPU, memory and disk use, its metadata and the problems on it and its services.

**Infrastructures** tracks the hosts in the environment, their health and their resource use. Engineers open it; Admins do not.

![Infrastructure list opened from a monitoring alert, its banner above two hosts marked with the CPU readings the alert caught](https://curade.ai/_astro/04-alt-infrastructure-landing.BBTkxOA2_1QCAwF.webp "Spot the hot host.")

## The list

Each card shows the host, its health and its headline resource figures. The health describes the host right now: widening the time window changes which hosts are in scope, never their health.

A host that is not healthy carries one plain line saying why. An unhealthy host names its open problem. A host whose health is **Unknown** says what monitoring recorded about it over the window, such as that it shut down, restarted or its monitoring agent stopped. When several hosts on screen share a reason, a line above the list says so.

On a custom range that ended in the past, every host reads as unknown, since health is only read live. Everything else on the page still covers the range you picked.

## The host's page

1. The header: the host's name and health, and its CPU, memory and disk averages, with **Copy details**.
2. **What's behind these numbers** ([Cluemetry](https://curade.ai/en/curade-mini/docs/features/cluemetry/)).
3. **Infrastructure Details**: the operating system, how the host is identified, the **Custom Metadata** your team sets on it, its management zones and its tags. A host in no zone reads **No management zones**.
4. The problems raised against the host inside the window.
5. The relationship map: the services on the host and what calls each of them ([Relationship map](https://curade.ai/en/curade-mini/docs/features/relationship-map/)).
6. **Problems on services running on this host**: what those services ran into over the same window.
7. Charts of CPU, memory and disk over the window.

## Hosts and their services

Monitoring raises a problem against the item it detected, so a slow service is named as the service, never as the host underneath it. That is why a host can carry no problems of its own while the services on it are struggling, and why the page lists their problems in a section of their own. A row that says its service also runs on other hosts belongs to a service that runs elsewhere too, so this host is only one of the places it could have struggled.

## Reading the charts

CPU, memory and disk are percentages over the window. The headline figure is the window's average while the chart plots each interval, so a calm average over a spiky chart is a real pattern worth a look. Tap a point to read that interval, and **Why?** explains it.