feat: Extended Prometheus metrics

[msalinas92]

Add additional Prometheus metrics to better observe internal cache behavior.

New metrics examples:

• cachebolt_evictions_total
• cachebolt_ttl_expired_total
• cachebolt_backend_errors_total{backend="s3"}

Acceptance Criteria:

• Metrics include appropriate labels (e.g., backend, URI).
• Available at /metrics without impacting performance.