> ## Documentation Index
> Fetch the complete documentation index at: https://docs.niro.cx/llms.txt
> Use this file to discover all available pages before exploring further.

# Environments

> Tag clusters as dev, staging, or production to control policy severity.

## What is a cluster environment?

Every cluster in Niro has an **environment** tag that tells Niro how to treat that cluster. The environment affects [policy](/features/policies) severity — checks that are informational on a dev cluster can escalate to high severity on a production cluster.

## Environment values

| Value         | Use case                                                                        |
| ------------- | ------------------------------------------------------------------------------- |
| `unspecified` | Default. No severity escalation. Use when you haven't classified a cluster yet. |
| `dev`         | Local development and testing clusters. Low severity on all findings.           |
| `staging`     | Pre-production validation. Moderate severity.                                   |
| `production`  | Customer-facing or business-critical clusters. Full policy severity escalation. |

## How environment affects policies

Some policy rules have different severity levels depending on the cluster environment:

| Policy rule               | Non-production severity | Production severity |
| ------------------------- | ----------------------- | ------------------- |
| Single replica            | Info                    | High                |
| Mutable `:latest` tag     | Low                     | High                |
| Missing resource requests | Medium                  | Medium              |
| Missing resource limits   | Medium                  | Medium              |
| Missing health probes     | Medium                  | Medium              |
| Privileged container      | High                    | High                |
| No rollback history       | Low                     | Medium              |

Production clusters get louder warnings on the risks that matter most in production: a single-replica deployment or a mutable image tag can cause downtime, and the policy engine reflects that.

<Note>
  Environment is declared by the operator — Niro doesn't try to infer it from namespace names or labels. This is intentional: an MSP might have a namespace called `production` in a dev cluster, or vice versa.
</Note>

## Setting the environment

### During cluster creation

When you add a new cluster, you can set the environment in the "Add cluster" dialog.

### After enrollment

Open the cluster detail page → **Settings** → change the **Environment** dropdown → save.

The change takes effect on the next heartbeat (within \~15 seconds). New policy findings will use the updated severity; existing findings are re-evaluated on the next heartbeat cycle.

## Environments and policy evaluation

Policy checks run at three points:

1. **Build time** — when you fill in the visual builder form. Findings appear inline as you edit.
2. **Merge time** — when a PR is opened against a linked repo. Findings appear as PR comments with the cluster's current environment in scope.
3. **Runtime** — continuously. Niro evaluates the live pod state against the cluster's current environment.

In all three cases, the environment tag is used to determine the final severity shown to you.

## Related

* [Policies & Standards](/features/policies) — full list of policy rules and how they're evaluated
* [Clusters](/concepts/clusters) — cluster lifecycle and settings