# Built-in metric views

Pre-built views for hosts, containers and common middleware, and how to add your own.

Last reviewed 2026-10-10

Where this page ends: charts from clicking metric names instead of writing PromQL, and your team's own queries added to the same catalogue so colleagues who do not write PromQL can use them too.

## What this page is for[​](#what-this-page-is-for "Direct link to What this page is for")

You have just connected a component, data is arriving, and you have no idea **which metrics are worth looking at**. Built-in metrics is that catalogue: organised by component type and category, each entry carrying a metric name, a unit, the PromQL and a plain-language description, and charting on one click.

A fresh install ships with more than 1300 entries covering Linux, MySQL, Redis, Nginx, Kafka, ClickHouse, Kubernetes and other common components.

Compared with the [Metrics explorer](https://n9e.github.io/docs/query/metric-explorer.md): that one is for people writing PromQL during an incident; this one is for routine checks and for finding out what a newly connected component even exposes.

**Explorer → Metrics → Built-in metrics** (`/metrics-built-in`).

## 1. Set a filter before clicking a metric[​](#1-set-a-filter-before-clicking-a-metric "Direct link to 1. Set a filter before clicking a metric")

![Built-in metrics](https://n9e.github.io/img/docs/query/metrics-built-in.webp)![Built-in metrics](https://n9e.github.io/img/docs/query/metrics-built-in.dark.webp)

The left column is **Filters**, and its job is to **attach a set of label conditions to the query whenever you click a metric**.

In the product's own words: with `{ident="n9e01"}` configured and selected, clicking `cpu_usage_idle` actually issues `cpu_usage_idle{ident="n9e01"}`.

**Set one up before you start clicking.** Without it, a metric like `cpu_usage_idle` returns every series in the environment — thousands of lines, and a stalled browser.

Click **+** at the top of the left column:

| Field             | What to put in it                                                       |
| ----------------- | ----------------------------------------------------------------------- |
| Name              | Something recognisable, e.g. "Trading production"                       |
| Data source       | Which source this filter applies to                                     |
| Filter conditions | The label conditions, e.g. `ident="n9e01"`                              |
| Authorized teams  | Which teams may use it, and whether **read and write** or **read only** |

## 2. Find the metric you want[​](#2-find-the-metric-you-want "Direct link to 2. Find the metric you want")

Four filters across the top: **Component type** (Linux, MySQL, Kubernetes…), **Category** (the second level, e.g. cpu / mem / disk under Linux), **Unit**, and a keyword search that matches the name, the PromQL and the description.

The table shows five columns by default: component type, category, name, unit, PromQL. **Display columns** at the top right adds expression type, metric type, custom fields, description and updated-by.

The **Name** column actually renders the metric's **description**, and it is a link; the real expression is in the PromQL column. Clicking it opens a query drawer that charts it straight away, and **Add panel** in that drawer stacks several metrics for comparison.

With rows selected, **More** offers export (JSON), import and delete.

## 3. Add one of your own[​](#3-add-one-of-your-own "Direct link to 3. Add one of your own")

Click **Add metric**:

| Field              | What to put in it                                                                   |
| ------------------ | ----------------------------------------------------------------------------------- |
| Component type     | Pick an existing one or create a new one                                            |
| Category           | The second level                                                                    |
| Expression type    | **Metric name** (just a name) or **PromQL** (a full expression with functions)      |
| Metric type        | Gauge / Counter / Histogram                                                         |
| PromQL             | The query itself                                                                    |
| Unit               | Used for display conversion                                                         |
| Custom fields      | Name plus value, for whatever context you want to carry                             |
| Metric description | Per language: **Name** (required) and **Description** — this is what the list shows |

Note: **there is no top-level name field**. The name lives inside the per-language description block, so a metric filled in only in Chinese will not show a name in the English UI.

Creating, editing and deleting are gated by the `/builtin-metrics/add`, `/put` and `/del` permission points.

## Where the catalogue comes from[​](#where-the-catalogue-comes-from "Direct link to Where the catalogue comes from")

This decides where you go to change something:

* The 1300-odd built-in entries come from `integrations/<Component>/metrics/*.json` in the backend and are **loaded when n9e-center starts**. No UI action triggers it, and there is nothing for you to "import";
* Metrics you create here are stored in the database. **For the same PromQL, the database row wins over the file** — so when a built-in entry cannot be edited, the workaround is to create a new one with the same PromQL;
* Integrations → Components has a **Metrics** tab under each component. That is another view onto the same data; editing either side is equivalent.

Two other places consume this catalogue: the **Built-in metrics** button beside the PromQL input in the Metrics explorer, and the [one-click drawer](https://n9e.github.io/docs/query/object-view.md) on the host list.

## Next[​](#next "Direct link to Next")

* Write the PromQL yourself: [Metrics explorer](https://n9e.github.io/docs/query/metric-explorer.md)
* Look at things host by host: [Object quick view](https://n9e.github.io/docs/query/object-view.md)
* What else a component template carries: [Import dashboards and rules](https://n9e.github.io/docs/data/integrations/import.md)