Monitoring IT systems involves tracking various metrics, such as CPU and RAM usage, which serve as foundational elements in a monitoring setup, supplemented by higher-level metrics that assess application performance from the user's perspective. A key challenge in monitoring is managing cardinality, the number of unique combinations of metric names and dimensions, which can exponentially increase as more dimensions are added, complicating data analysis. Time-series databases (TSDBs) used for storing metrics face unique challenges in balancing data ingestion and retrieval efficiency, while monitoring practices must also adapt to the demands of microservices and Docker-based workloads, where an additional layer of monitoring for container health is necessary. High-level performance metrics, while versatile, present their own cardinality challenges, necessitating careful selection of dimensions to ensure meaningful insights without overwhelming the system. Effective monitoring requires managing stale data, devising lifecycle policies, and choosing appropriate tools based on workload complexity, striking a balance between detailed dimension use and data clarity. Success in this area involves designing a monitoring system that aligns with organizational needs and adapts over time, leveraging advanced tools to provide context and clarity in data analysis.