OpenObserve

Use it when

• You have no unified view of logs, metrics, and traces and want a self-hosted OpenTelemetry-native platform instead of paying for Datadog or Splunk
• Elasticsearch-based log storage costs and cluster management overhead are too high and you want a single binary with S3-backed columnar storage

What it solves

Not the fit when

• Workloads needing to modify or delete individual ingested records; data is immutable and removal operates through retention boundaries.
• SSO, advanced RBAC, audit trails, federated cross-cluster search, and sensitive-data redaction without the Enterprise features documented by the project.
• Exposing the local HTTP example to a network without TLS, unique credentials, access controls, backups, and a reviewed retention policy.
• Treating the advertised storage-cost comparison as universal; the up-to-140x claim is workload- and configuration-specific.
• Deployments that cannot satisfy AGPL-3.0 obligations.
• frontend visual regression testing
• test suite regression detection
• application performance fixes themselves

Install

Set unique high-entropy ZO_ROOT_USER_EMAIL and ZO_ROOT_USER_PASSWORD values outside shell history, mount durable storage, then run the documented public.ecr.aws/zinclabs/openobserve container on port 5080. Do not reuse the README's example credentials.

Invoke

For local evaluation, open http://localhost:5080 and sign in with the credentials you set; ingest logs, metrics, and traces over OpenTelemetry OTLP, query logs and traces with SQL and metrics with SQL or PromQL, and build dashboards, alerts, and ingest pipelines. Add TLS and access controls before any network exposure.

Alternatives

No reviewed alternatives recorded yet.