How can latency that appears only under concurrent requests be investigated in discloud when using environment variable configuration?

When using discloud’s environment‑variable configuration (similar to Azure App Service app settings), latency sometimes increases only when the application handles multiple concurrent requests. The goal is to identify whether the delay originates from configuration loading, container startup, or request‑processing paths under load.

What specific aspects of discloud’s concurrent request handling should be examined to isolate this latency, and which diagnostic signals (e.g., metrics, logs) would help confirm the root cause?