[Feature] <title>[Feature] Add AI-based anomaly explanation to alerts

### Feature Request

Add a new feature to improve AI-based monitoring by providing automatic anomaly explanations for alerts.

### Is your feature request related to a problem? Please describe

Alerts in HertzBeat only show a generic warning message. Users need to manually analyze metrics and logs to understand the root cause.
This feature would automatically generate a clear explanation of why the alert happened.

### Describe the solution you'd like

Add an AI-powered explanation feature for alerts.

When an alert triggers, HertzBeat should:
1. Collect relevant metrics and logs for the alert time range.
2. Send them to the AI module using SpringAI.
3. Generate a short human-readable explanation of the cause.
4. Display the explanation in the alert detail page.

The explanation should include:
- Possible root cause
- Suggested troubleshooting steps
- Related metrics or logs

### Describe alternatives you've considered

1. Manually adding more detailed alert messages: This requires a lot of manual effort for each alert type and cannot scale.
2. Using static rule-based explanations: This is inflexible and cannot handle complex patterns in monitoring data.
3. Adding only metric graphs without explanations: Users still need to analyze the data manually, which is time-consuming.

### Additional context

This feature is inspired by the need for faster troubleshooting in observability systems.
Other monitoring tools provide root cause analysis using AI, but HertzBeat currently only shows alerts without explanations.
The AI explanation feature can be implemented using SpringAI and can be optional (configurable) for users who want it.

The implementation can be added as a new module or plugin to keep the core system clean.

The issue names no files, tests, or entry points. Start by locating HertzBeat's alert detail page, alert data collection, and AI module, then review how SpringAI could receive metrics and logs. Done means an optional alert explanation displays a possible cause, troubleshooting steps, and related data.

Written by the indexing model from the issue text.