Avoid repeated YAML config loading per request

[lbedogni]

Problem
Functions load_network_delay_config(), load_local_inference_config(), and load_verbose_config() each open and parse settings.yaml. This happens per request.

Impact

• Unnecessary CPU and I/O overhead per request.

Proposed Fix

• Load settings.yaml once at server startup (module level or in RequestHandler.__init__) and reuse the config dict.
• If dynamic reload is needed, implement a SIGHUP handler or admin endpoint.

Files: src/server/communication/request_handler.py Labels: performance

[lbedogni]

✅ RESOLVED

The logger has been centralized in src/server/logger/ with auto-configuration from settings.yaml.

Implementation details:

• Centralized logger in src/server/logger/log.py
• Auto-configuration via configure_logger_from_settings() at import time
• verbose flag in settings.yaml controls log levels efficiently
• When verbose: false, logger level raised to WARNING to short-circuit all debug/info formatting

This issue is now complete and can be closed.

[lbedogni]

✅ Completed - see AGILE_ACTION_PLAN.md