fix: Get true lock_timeout and statement_timeout

pgwatch sets lock_timeout to 100ms during metric collection, which causes the monitoring to report 100ms instead of the actual configured value.

the same applies to statement_timeout.

This fix queries reset_val from pg_settings for lock_timeout and statement_timeout to get the genuine configuration value.

Closes #61 (closed)

---

Note

Report the actual lock_timeout by using reset_val in the settings metric and add unit/integration tests to verify it.

• Metrics config (config/pgwatch-prometheus/metrics.yml):
  • settings: Use reset_val instead of setting for lock_timeout (documented in description) and apply the same CASE logic for numeric_value.
• Tests:
  • Add tests/settings/test_settings_metric.py with unit tests ensuring the SQL/description changes and an integration test validating lock_timeout returns the reset value despite session overrides.

^{Written by Cursor Bugbot for commit 7891edca. Configure here.}

#61 (closed)