Developed and integrated a Prometheus Metrics Recorder Periodic Upkeep Task for the tokio-rs/axum repository, focusing on improving the reliability and hygiene of observability data. The solution leveraged Rust and asynchronous programming with Tokio to implement a background task that periodically invokes upkeep on a cloned metrics recorder handle. This approach addressed the challenge of metric drift and the accumulation of stale data by automating cleanup within the metrics subsystem. The work demonstrated practical use of background tasks and system monitoring, and included clear example-based guidance to help users adopt Prometheus metrics upkeep patterns in their own axum-based applications.