Timbala is in a very early stage of development and is not yet production-ready. Please do not use it yet for any data that you care about.

There are several known issues that prevent any serious use of Timbala at this stage; please see the MVP milestone for details.

Timbala is a distributed, fault-tolerant time-series database intended to provide durable long-term storage for multi-dimensional metrics.

It is designed to integrate easily with Prometheus, supports PromQL and is API-compatible with Prometheus, but can be used standalone.

Data stored in Timbala can be visualised using Grafana by configuring a Prometheus data source pointing to Timbala.

• one server binary
• no external dependencies
• all nodes have the same role
• all nodes can serve read and write requests

• no single points of failure
• data is replicated and sharded across multiple nodes
• planned features for read repair and active anti-entropy

• high write throughput and availability