OpenData Timeseries Brings Object-Store-Native Architecture to Prometheus Ecosystem
Key Takeaways
- ▸OpenData Timeseries applies object-store-native design principles (successfully validated by WarpStream and turbopuffer in other database categories) to Prometheus-compatible timeseries databases for the first time
- ▸The system dramatically simplifies operational complexity by eliminating stateful storage servers, quorum replication, and manual sharding/rebalancing—reducing the typical Cortex architecture from six required services to a minimal two-tier design
- ▸Full compatibility with Prometheus tooling (PromQL, remote write, Grafana) and OpenTelemetry metrics enables adoption without ecosystem changes
Summary
OpenData has released OpenData Timeseries, an MIT-licensed, Prometheus-compatible timeseries database built on SlateDB, an object-store-native LSM tree. The system supports PromQL, Prometheus scraping, Prometheus remote write, and OTLP metrics ingest, bringing the operational simplicity and cost efficiency of object-store-native architectures to observability stacks. Unlike traditional Prometheus-compatible systems like Cortex and VictoriaMetrics that require complex distributed storage management with multiple stateful services, OpenData Timeseries simplifies operations by relegating disks to caches and using object storage as the sole durable persistence layer. This architectural shift eliminates the need for quorum replication, complex sharding management, and costly inter-AZ networking, potentially reducing storage costs by 80-90% compared to disk-based systems while maintaining compatibility with the Prometheus and Grafana ecosystem.
- Cost structure mirrors object-store-native databases with 80-90% storage savings and elimination of expensive inter-AZ replication networking costs
Editorial Opinion
OpenData Timeseries addresses a genuine gap in the observability landscape. While object-store-native designs have proven their worth in streaming and vector databases, observability teams have been locked into operating expensive distributed systems or paying premium hosted solutions. This project could significantly democratize enterprise-scale observability by combining the operational simplicity of stateless compute with the cost efficiency of cloud object storage—though success will depend on real-world performance under high-cardinality metric workloads.
