Promscale is a deprecated unified observability backend written in Go that integrated metric and trace storage capabilities for Prometheus, Jaeger, and OpenTelemetry systems. Built on PostgreSQL and TimescaleDB, it provided a consolidated platform for organizations seeking to manage both metrics and distributed traces through a single backend system. The project is no longer maintained, as indicated by its deprecation status and the warning in its README directing users to issue 1836 for more information about the discontinuation.

The repository was designed to address a key gap in observability infrastructure by eliminating the need for separate storage systems for different telemetry types. Promscale functioned as a 100% PromQL-compliant Prometheus remote storage backend while simultaneously serving as a certified Jaeger storage backend. This dual capability meant organizations could ingest metrics through Prometheus remote write protocols and traces through Jaeger collectors or the OpenTelemetry Protocol, storing both in the same PostgreSQL-based system. The architecture consisted of just two components: the Promscale Connector and the Promscale Database, intentionally kept simple to reduce operational complexity compared to alternatives like Elasticsearch or Cassandra for trace storage.

For Prometheus users, Promscale offered centralized metric storage across multiple Kubernetes clusters with multi-tenancy support, enabling single-pane-of-glass visibility. It supported PromQL alerting rules, recording rules for downsampling, and per-metric retention policies to optimize storage costs and query performance for long-term trend analysis. The system could scale to millions of series and hundreds of thousands of samples per second on a single PostgreSQL node through TimescaleDB's optimization capabilities.

For Jaeger and OpenTelemetry users, Promscale provided durable and scalable trace storage with native OTLP support. A distinctive feature was the ability to query traces using SQL, enabling deeper trace analysis beyond Jaeger's built-in filtering capabilities. The system supported Service Performance Management features in Jaeger and included an Application Performance Management experience in Grafana with customizable dashboards built on SQL queries against trace data.

The repository's GitGenius classification data reveals it was categorized across multiple domains including scalability, real-time analysis, SQL extensions, analytics, PostgreSQL extensions, horizontal scaling, data storage, observability, cloud-native data ingestion, monitoring, query optimization, and time-series database functionality. The project maintained connections with related repositories including timescale/timescaledb, netdata/netdata, and rust-lang/rust through overlapping contributors. However, issue and pull request response latency data shows a median response time of over 24,000 hours, reflecting the project's deprecated status and lack of active maintenance. The most actively tracked issue label was Deprecation, with one item recorded, underscoring the project's end-of-life state.