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

# Monitoring as Code with AI

> Use AI assistants to generate and maintain DevHelm monitoring configurations

AI assistants can generate `devhelm.yml` files, suggest assertions, and help you maintain monitoring configs as your infrastructure evolves.

## Generating configs from scratch

Describe your infrastructure and let the AI create a monitoring config:

> "I have a REST API at api.example.com with /health, /v1/users, and /v1/orders endpoints. Create a DevHelm config that monitors all three with appropriate assertions."

An AI assistant with access to DevHelm's MCP tools or YAML schema knowledge can produce:

```yaml theme={null}
version: "1"
monitors:
  - name: API Health
    type: HTTP
    config:
      url: https://api.example.com/health
      method: GET
    frequencySeconds: 30
    assertions:
      - config:
          type: status_code
          expected: "200"
          operator: equals
      - config:
          type: response_time
          thresholdMs: 2000

  - name: Users API
    type: HTTP
    config:
      url: https://api.example.com/v1/users
      method: GET
    frequencySeconds: 60
    assertions:
      - config:
          type: status_code
          expected: "200"
          operator: equals

  - name: Orders API
    type: HTTP
    config:
      url: https://api.example.com/v1/orders
      method: GET
    frequencySeconds: 60
    assertions:
      - config:
          type: status_code
          expected: "200"
          operator: equals
```

## Reviewing and improving existing configs

Share your current `devhelm.yml` and ask for improvements:

> "Review this config. Are there any gaps in monitoring coverage or assertions that could reduce false positives?"

The AI can identify:

* **Missing assertions** — endpoints without response time checks
* **Inconsistent frequencies** — critical and non-critical services at the same interval
* **Missing alerting** — monitors without notification policies
* **Coverage gaps** — services mentioned in code but not monitored

## Maintaining configs as code evolves

When you add new API endpoints, the AI can update your monitoring config:

> "I just added a /v1/payments endpoint. Add a monitor with the same pattern as the existing ones, but check every 30 seconds since it's payment-critical."

This keeps monitoring in sync with your codebase without manual editing.

## Best practices for AI-assisted config

### Always validate

Run `devhelm validate` on any AI-generated config before deploying:

```bash theme={null}
devhelm validate devhelm.yml
```

AI models can produce syntactically valid YAML that uses wrong field names or invalid enum values. Validation catches these.

### Use plan before deploy

Preview changes before applying them:

```bash theme={null}
devhelm plan -f devhelm.yml
```

Review the plan output to confirm the AI made the changes you expected.

### Keep humans in the loop

AI-generated configs are a starting point. Review:

* Are the assertions appropriate for each endpoint?
* Are the frequencies correct for the service criticality?
* Are secrets referenced correctly (not hardcoded)?
* Are the right alert channels attached?

### Version control everything

Commit AI-generated configs to Git like any other code change. This gives you review through PRs and rollback through Git history.

## Workflow: AI agent + MaC

The most powerful pattern combines AI agents with monitoring-as-code workflows:

1. **Agent generates** config based on your description
2. **You review** the generated YAML in a PR
3. **CI validates** with `devhelm validate`
4. **CI previews** with `devhelm plan --detailed-exitcode`
5. **You approve** the PR
6. **CI deploys** with `devhelm deploy --yes`

The AI handles the tedious config generation; your CI/CD pipeline handles safe deployment.

<CardGroup cols={2}>
  <Card title="YAML file format" icon="file-code" href="/mac/yaml/file-format">
    Complete YAML schema reference.
  </Card>

  <Card title="MCP Server" icon="server" href="/sdk/mcp/overview">
    Connect your AI agent to DevHelm.
  </Card>
</CardGroup>