Observability is the ability to measure a system's state based on the data it generates. To be effective, observability tools must ingest data from various sources, typically in the form of metrics, traces, logs, and metadata. Promscale is an easy-to-use observability backend built on top of TimescaleDB, enabling engineers to store all their observability data in a single mature and scalable store and analyze data through a unified SQL interface. Earlier this month, Promscale launched support for traces, allowing developers to interrogate trace data and unlock new insights. Prometheus downsampling leverages continuous aggregates, reducing the resolution of data while its size, making it cheaper to store and query. Metric monitoring is crucial in micro-service-based systems running on Kubernetes, where Prometheus excels due to its auto-discovery mechanism and support for dimensional metrics. However, Prometheus metric data can be overwhelming, with hundreds of thousands of samples per second. Downsampling using recording rules or Promscale's continuous aggregates helps reduce storage costs and improve query performance by summarizing individual data points into aggregate statistics. Continuous aggregates offer timeliness, flexibility, and improved query capabilities compared to recording rules. To use continuous aggregates, engineers need to create a materialized view with a specific schema and configure the aggregation function. The benefits of using Promscale's continuous aggregates include faster queries, reduced storage costs, and improved data retention for downsampled data. However, there are considerations when choosing between recording rules and continuous aggregates, including access to recent data, size of the time-bucket, number of metrics in materialization, and query flexibility.