OpenObserve

OpenObserve

WHAT IS IT?

OpenObserve is a self-hostable observability platform that brings logs, metrics, traces, RUM, dashboards and alerts together in one place. Written in Rust, it positions itself as an alternative to Datadog, Splunk or Elasticsearch, with a heavy focus on storage cost. It runs as a single binary, from standalone mode all the way to HA clusters.

WHY IS IT INTERESTING?

  • Object storage, built for cost: data is written as columnar Parquet straight to S3 (or compatible), with intelligent caching. The project claims up to 140x cheaper than Elasticsearch.
  • Single binary: start standalone to try it, then move to a distributed cluster without switching tools.
  • OpenTelemetry-native: OTLP ingestion with no proprietary agent, so no lock-in on the collection side.
  • SQL queries: SQL for logs and traces, SQL or PromQL for metrics, no in-house language to learn.
  • Full platform: web UI with a drag-and-drop dashboard builder, alerting across logs/metrics/traces and session replay on the RUM side.

USE CASES

  • Centralize an infra's logs without blowing up the storage bill.
  • Follow distributed traces across a microservices architecture.
  • Replace an ELK stack that has become heavy to operate.
  • Plug in your metrics and keep PromQL without running a full Prometheus alongside.