OpenObserve is an open source observability platform for logs, metrics, and traces — a self-hosted Datadog and Splunk alternative, built in Rust, that stores telemetry as compressed columnar files on object storage to keep costs low at scale.

What is OpenObserve?

OpenObserve (O2) is an open source observability platform that collects logs, metrics, and traces in one place. Built in Rust, it stores telemetry as compressed Parquet columnar files on S3-compatible object storage, so you can search events, build dashboards, and alert on your systems without the storage bill of an Elasticsearch or Splunk stack. It’s one of the more cost-focused open source monitoring and observability tools.

What is OpenObserve best for?

OpenObserve is best for teams that want a single tool for all three observability signals — logs, metrics, and traces — instead of stitching together the Grafana LGTM stack or paying Datadog and Splunk per host. Its object-storage design suits high-volume telemetry where storage cost and long retention matter, and native OpenTelemetry support makes it a natural fit for OTel-instrumented applications.

What can OpenObserve do?

Ingest and search logs with full-text search, SQL queries, quick filters, and a visual query builder
Store metrics and query them with SQL or PromQL across 19+ chart types
Capture distributed traces via OpenTelemetry, with waterfalls, flame graphs, and Gantt views
Build drag-and-drop dashboards with template variables
Real User Monitoring (RUM) with session replay, Core Web Vitals, and error tracking
Alerting — threshold, scheduled, real-time, and anomaly detection — plus incident correlation and lifecycle tracking
Pipelines that enrich, redact, and normalize data at ingestion
Run as a single Rust binary for small setups, or a clustered high-availability deployment for petabyte scale

Where does OpenObserve fall short?

Data is immutable once ingested. You can’t edit or delete individual records — corrections or targeted deletions mean dropping an entire retention period, not a single row, which complicates compliance workflows like GDPR erasure.
Some features sit behind the Enterprise edition. SSO, RBAC, and AI-observability tooling live in the Enterprise build, which is free to self-host only up to 50 GB/day of ingestion; past that you need a paid license, and the pure AGPL edition omits them.
Younger ecosystem than Elasticsearch or Grafana. There are fewer prebuilt integrations, community dashboards, and third-party plugins, so you lean on OpenTelemetry and build more yourself rather than pulling from a large plugin marketplace.

Is OpenObserve free?

Yes — the open source edition (AGPL-3.0) is free to self-host forever, and the self-hosted Enterprise edition, which adds SSO and RBAC, is free up to 50 GB/day of ingestion. Beyond that, or on the managed OpenObserve Cloud, pricing is usage-based pay-as-you-go: roughly $0.50 per GB ingested and $0.01 per GB queried, with unlimited users and no per-seat charges.

What does OpenObserve replace?

OpenObserve is built as a cheaper, self-hostable stand-in for hosted observability suites. It’s a Datadog alternative and a Splunk alternative for logs, metrics, and traces, and it directly targets Elasticsearch/ELK-based logging stacks — the project claims roughly 140x lower storage cost and about a quarter of the hardware versus Elasticsearch. Teams already running Grafana sometimes adopt it to consolidate ingestion, storage, and querying into one component instead of four.

FAQ

Is OpenObserve open source? Yes. The core is licensed under AGPL-3.0 and developed in the open on GitHub. A separate Enterprise edition adds features such as SSO and RBAC and is free to self-host up to 50 GB/day of ingestion.

Can I self-host OpenObserve for free? Yes. Run the open source edition on your own infrastructure at no cost — it ships as a single Rust binary and can store data on local disk or any S3-compatible object storage like MinIO, AWS S3, or GCS.

Is OpenObserve a good Datadog or Splunk alternative? For teams that want to control cost and own their data, yes. It covers logs, metrics, and traces in one platform with usage-based pricing instead of per-host or per-seat billing, though its integration ecosystem is smaller than Datadog’s.

What do I need to run OpenObserve? A single binary and local disk are enough for small or single-node setups. For production high availability, you add S3-compatible object storage plus a metadata database such as Postgres or MySQL — there’s no hard dependency on Elasticsearch or Kafka.