System of record = Elasticsearch
Product state, quality results, lineage events, pathway measurements, and incidents live in Elasticsearch. Kibana remains the deep investigation surface.
Product
Elasticsearch-native data observability
DataObs is built for teams that already run Elasticsearch and want deeper visibility into data quality, pipelines, lineage, streams, cost, business impact, and AI agents — without leaving their existing analytical and investigation platform.
What makes DataObs different
Open standards first
OpenTelemetry for metrics, logs, and traces. OpenLineage for job and lineage events. No proprietary agent lock-in for core signals.
Six clear pillars
Platform, Data Pipeline, Data, FinOps, Business, and AI/Agent — designed so every capability maps to a real operational question.
Self-managed & controllable
Run it on your own Elasticsearch cluster. Full visibility into storage, retention, and cost. Optional delivery support when you need it.
Core product surfaces
Freshness, nullness, uniqueness, schema, row-count, distribution drift, and more.
Dataset and column lineage with upstream root-cause and downstream blast radius.
Job runs, lag, throughput, pathway latency, and health.
Alerts, cases, evidence, and guided remediation paths.
Command Center, Assets, Pathways, Streams, and incident views (in active development).
Current status
DataObs is under active productisation. Capability state is tracked in an evidence-gated ledger. The repository is not yet a production release, but core quality, lineage, pipeline, and Elasticsearch storage foundations are progressing rapidly.
See the main repository for the latest capability ledger, architecture docs, and roadmap.