Add debug logging for trace metric API calls when debug mode is enabled

After landing the #997 split, add explicit debug-level logging around metric capture calls so metrics behavior is easier to inspect when SDK debug mode is on. This follows the SHOULD-level diagnostics guidance in the [trace metrics spec](https://develop.sentry.dev/sdk/telemetry/metrics/).

### Scope from #997
No direct code extraction is needed from [#997](https://github.com/getsentry/sentry-rust/pull/997); this is a follow-up gap identified during spec review.

### Details
Add targeted, low-noise debug logging in the metric capture path (for example around drop/disable decisions and enqueue/flush boundaries) that mirrors existing SDK debug conventions. The goal is to improve observability for SDK integrators without changing functional behavior.