OpenData Timeseries Launches MIT-Licensed Prometheus-Compatible Database on Object Storage
Key Takeaways
- ▸OpenData Timeseries is a Prometheus-compatible TSDB built on object-store-native architecture, offering 80-90% cost savings compared to managed observability platforms
- ▸The system eliminates operational complexity inherent in traditional distributed TSDBs by using object storage as the only durable persistence layer, allowing compute nodes to remain stateless
- ▸Internal benchmarks demonstrate the ability to serve 4.7 billion samples per day across 3.3 million active series for approximately $500/month in compute costs
Summary
OpenData has announced OpenData Timeseries, a new MIT-licensed, Prometheus-compatible timeseries database built on SlateDB, an object-store-native LSM tree architecture. The system supports PromQL, Prometheus scraping, Prometheus remote write, and OTLP metrics ingest, enabling users to power Grafana dashboards and alerts at significantly lower costs than managed observability providers. According to internal benchmarks, the platform can serve 4.7 billion samples per day across 3.3 million active series for approximately $500 per month in compute costs.
The key innovation behind OpenData Timeseries is its object-store-native design, which simplifies operational complexity by eliminating the need for stateful distributed storage systems. Unlike traditional Prometheus-compatible databases like Cortex or VictoriaMetrics that require multiple stateful components for durability and replication, OpenData Timeseries uses object storage as its only persistent layer, allowing compute to remain stateless. This architectural approach has already proven successful in other domains, with projects like WarpStream (streaming) and turbopuffer (vector databases) demonstrating superior performance and cost efficiency.
By moving away from complex sharding, replication, and rebalancing requirements, OpenData Timeseries dramatically reduces operational overhead. The system can easily scale by adding or removing nodes without state shuffling, and it leverages the 80-90% cost savings of cold data storage on object storage versus traditional disk-based systems. The project fills a notable gap in the observability stack, where modern alternatives had not yet adopted object-store-native architectures despite their proven benefits.
- The MIT license and full PromQL compatibility enable drop-in replacement capabilities for existing Prometheus-based observability stacks and Grafana deployments
Editorial Opinion
OpenData Timeseries represents an important maturation of observability infrastructure by bringing proven object-store-native database patterns to the timeseries domain. The dramatic cost reduction—from expensive managed observability platforms to $500/month for substantial workloads—could democratize enterprise-grade observability for mid-market and cost-conscious organizations. While the architectural simplification is compelling, successful adoption will depend on community validation of performance characteristics, operational stability at scale, and ecosystem integration with the broader observability tooling landscape.
