Overview

Metric View = platform built-in / user-defined metric catalog, organizing common metrics by “component type + category”, so people unfamiliar with PromQL can click a metric name and see the chart.

Sidebar path: Data Query → Metrics → Metric View Tab, URL /metrics-built-in.

Use cases:

• A new component like MySQL/Redis/Nginx was added, and you don’t know which metrics are worth looking at;
• Someone on the team isn’t familiar with PromQL and wants “out-of-the-box”;
• Standardize the “ops inspection” action: register the daily must-see metrics in the catalog, and go through them one by one during inspection.

Difference from “Ad-hoc Query”:

Dimension	Ad-hoc Query	Metric View
Input	Write PromQL yourself	Click predefined metric name
Audience	Engineers familiar with PromQL	Everyone (including non-coders)
Suitable for	Ad-hoc troubleshooting, re-computation	Routine inspection, new component exploration

Page Layout

The page is split left and right:

Left: Filter Conditions

Defines a global label filter: all metrics, when clicked for charting, auto-carry these label conditions.

For example: add ident="n9e01" in filters, and clicking cpu_usage_idle issues the actual PromQL cpu_usage_idle{ident="n9e01"} — only the CPU of host n9e01, avoiding thousands of lines at once.

Strongly recommend setting filters first before clicking metrics, or high-cardinality metrics may crash the browser.

Right: Metric Catalog

Control	Description
Component type	Filter by component category, e.g. Linux, MySQL, Redis, Nginx, Kafka and other built-in components
Category	Secondary category, e.g. under Linux: CPU / Memory / Disk / Network
Unit	Filter by metric unit (percentage / bytes / count, etc.)
Keyword search	Fuzzy search in metric name / PromQL / notes

List fields:

Column	Meaning
Component type	Which component the metric belongs to
Category	Secondary category
Metric name	Click to open the Graph view directly
Unit	Used for unit conversion in display
PromQL	Actual query for the metric (some wrapped in rate(), etc.)
Actions	Edit / Delete / Copy (custom metrics editable)

Built-in vs Custom Metrics

The “Create Metric” button at top right adds custom metrics to put team-specific queries into the catalog. Fields:

• Component type, category (pick existing or new)
• Metric name / notes
• PromQL (actual query)
• Unit, axis type, and other rendering parameters

Built-in metrics are imported from Integration Center → Template Center — if you need to add a batch of metrics (e.g. you added Tomcat monitoring), go to the Template Center and import the matching integration pack — no need to manually create them one by one.

Hands-on: Turn “Daily Inspection” into a SOP

Classic usage: register the daily/weekly inspection metrics into the catalog, forming a “click-and-go checklist”:

1. List inspection metrics: with the team’s Tech Lead, list metrics to scan daily/weekly, categorized by business/component;
2. Bulk import: import from Template Center where possible; for those that can’t, manually use “Create Metric” to add;
3. Set filters: configure left filters by current instances/clusters of interest;
4. Click through: review each metric’s Graph; screenshot anomalies for weekly reports.

FAQ

Q1: No metrics in the list — what to do?

A: The Metric View in this environment is associated by data source. Check:

1. Is the top data source selected? (data sources selected on “Ad-hoc Query” Tab pass through);
2. Is the data source’s associated template imported? Go to Template Center and find the component’s “metric template” to import;
3. Or manually click “Create Metric” to add.

Q2: Why is there no Graph after clicking the metric name?

A: Usually the PromQL doesn’t return data on the current data source.

• Check if the left “filter conditions” have wrong labels (e.g. ident="nonexistent-host");
• Copy the metric’s PromQL directly to Ad-hoc Query and run to verify.

Q3: Built-in metric’s PromQL doesn’t fit my needs — can it be modified?

A: Built-in metrics themselves cannot be modified directly (controlled by templates). Two ways:

• Copy and modify: “clone” it as a custom metric, rename, modify the PromQL to what you want;
• Modify the template: go to Template Center, modify the corresponding “metric template”, and re-import — all instances of that component type update accordingly.

References

• Ad-hoc Query
• Quick View
• Template Center (import the metric pack of built-in components)