feat(metric): add get_service_metrics tool

[matty69v]

Hey team,

Continuing my contributions after #209 (force_check) and #210 (get_resource_timeline). Still dogfooding the MCP server with Claude Code against a staging Centreon.

The gap

After using list_resources to confirm a service is in WARNING or CRITICAL state, the natural follow-up is "what's the actual value?" The output string sometimes contains metrics, but it's unstructured text that the LLM has to parse heuristically. Meanwhile, the structured metric data (current value, unit, thresholds) is already available via the API but not exposed through MCP.

What this PR adds

A new get_service_metrics tool in a dedicated components/metric.py module:

async def get_service_metrics(host_id: int, service_id: int) -> list[Metric]

• Returns all metrics for a service with their current value, unit, and warning/critical thresholds
• Tagged readOnlyHint=True, idempotentHint=True
• Endpoint: GET /monitoring/hosts/{host_id}/services/{service_id}/metrics (API docs)

Implementation notes

• Metric model in utils/type.py with a list(host_id, service_id) static method
• Dedicated components/metric.py module (one module per API doc section, per reviewer feedback on feat(resource): add force_check tool to trigger checks on demand #209/feat(resource): add get_resource_timeline tool #210)
• Scalar parameters (host_id, service_id) rather than a wrapper object

Tool count

12 → 13. README updated.

Tests

2 new tests, all passing alongside the existing suite:

• tests/utils/test_type.py::test_metric_list
• tests/components/test_metric.py::test_get_service_metrics

33 passed in 1.87s
ruff check .          All checks passed!
mypy centreon_mcp     Success: no issues found in 19 source files

No new dependencies, no breaking changes.

Summary by Aikido

Security Issues: 0	🔍 Quality Issues: 2	Resolved Issues: 0

🚀 New Features

• Added get_service_metrics tool with Metric model and tests

^{More info}

[matty69v]

Hello team do we have some news from my 2 pr's ?