The open-source observability database
A columnar engine with object storage as primary storage. Ingest through OpenTelemetry and Prometheus Remote Write, query observability data with SQL and metrics with PromQL.
THREE SYSTEMS
Bring your own protocols
Ingest through OpenTelemetry, Prometheus Remote Write, Loki Push, and Elasticsearch Bulk. Query observability data with SQL and metrics with PromQL. Migrate ingestion one signal at a time without rebuilding your collectors.
Why now: agents
Agents are changing both sides of observability. Agentic applications emit high-cardinality telemetry. Investigation agents explore that same data in parallel — a different concurrency profile than a dashboard ever produced. Both need the same thing underneath: one place to write it and one place to query it.
Human
Each step waits for a result before the next one is chosen.
Correlation happens in the engineer's head.
Agent
Many queries at once. Branches that miss get dropped, the ones that hit go deeper.
Branches abandoned halfway
Li Auto
SGCC
TensorFusion
Li AutoSGCCTensorFusionLi AutoSGCCTensorFusion
Recent data is served from memory and local-disk caches; long retention lives on object storage — same engine, same SQL
Retention policies, downsampling, and continuous aggregation are database features, so derived metrics do not need a second pipeline
Inverted, skipping, and fulltext indexes; distributed tables partition by column ranges
Object storage as primary — S3, GCS, Azure Blob, and S3-compatible endpoints
Up to 50× lower storage cost through compression and object storage
OceanBase Cloud: 300 TB of logs, 60%+ lower storage cost after migrating off Loki
Compute and storage are disaggregated
One binary, from a single node to a 100-node Kubernetes cluster
Enterprise adds read replicas, workload isolation, automated repartitioning, and region load balancing
Plan That Fits Your Needs
Enterprise User?
Contact us
One columnar engine for metrics, logs, and traces, backed by object storage. SQL across observability data, PromQL for metrics. Apache-2.0 licensed core.
Production-ready standalone or cluster deployment
Retention, downsampling, and continuous aggregation built in
Community-driven, open governance
Scale observability on your infrastructure — workload isolation, independent read capacity, automated repartitioning and region load balancing, HA, and support.
Mission-critical systems monitoring and analytics
Higher performance and data governance
Industries requiring high security and compliance (finance, healthcare etc.)
Organizations needing automated operations and intelligent resource scheduling
Teams that need data to stay in their own environment, with enterprise support
