VictoriaLogs is an Apache-2.0, Go-based log database from VictoriaMetrics that stores logs without a fixed schema and is built to handle terabytes of log data per day, aimed at teams that run their own logging infrastructure.
What it is
VictoriaLogs is a high-performance, lightweight, zero-config, schema-free database for logs, distributed under the Apache-2.0 licence and written in Go. It ships in two forms: a single-node version and a cluster version, and both are open source and free to use. It scales vertically and horizontally, from very small setups to large-scale deployments handling terabytes of logs per day.
The concrete problem it solves is storing and querying large volumes of log data without the schema design, tuning, and operational weight that a traditional search-oriented log store demands. It lives in the logging and observability ecosystem alongside tools such as Elasticsearch and Grafana Loki, both of which appear in its topic list. For people coming from those systems, VictoriaLogs provides a query language called LogsQL, and the project publishes a "LogQL to LogsQL" playground and an "SQL to LogsQL" playground to translate queries from the query languages those other tools use.
Key capabilities
- Single-node and cluster deployments are both open source and free, so scaling out does not mean moving to a paid tier.
- Schema-free, zero-config design for logs, with throughput described as terabytes per day.
- LogsQL query language, with "SQL to LogsQL" and "LogQL to LogsQL" conversion playgrounds.
- Built-in web UI, available as a hosted playground.
- Grafana plugin for VictoriaLogs, with a hosted Grafana playground for Kubernetes logs.
- Distribution as binary releases, plus Docker images
victoriametrics/victoria-logs on Docker Hub and Quay.
- Documented upgrade path and a public CHANGELOG for version tracking.
Who uses it and how
- Small setups running the single-node version, up to large-scale deployments handling terabytes of logs per day.
- Kubernetes environments collecting cluster logs, as indicated by the Kubernetes topic and the Kubernetes-logs Grafana playground.
- Grafana users who query VictoriaLogs through the dedicated Grafana plugin rather than through another log backend.
- Observability and OpenTelemetry pipelines, where it serves as the log storage layer.
- SIEM-style workloads, where long-retention log search and investigation are the primary task.
Getting started
Install from binary releases on GitHub, or run the Docker image victoriametrics/victoria-logs from Docker Hub or Quay. The documentation separates the single-node version and the cluster version, and includes a documented upgrade procedure.
How it compares
The facts name Elasticsearch and Grafana Loki as adjacent tools in the same space, and the project's LogQL-to-LogsQL and SQL-to-LogsQL playgrounds show that it expects users arriving from Loki and SQL-style query backgrounds. Its distinguishing claims as stated are the open-source single-node and cluster versions and the zero-config, schema-free model.
When to use it — and when not to
Choose it when a self-hosted, Apache-2.0 log store that scales from one node to a cluster is acceptable, and when the team is prepared to run either the single-node or the cluster deployment itself; the facts list no hosted or managed offering, only documentation and playgrounds. Teams that require a vendor-operated service should look elsewhere. Also weigh the project's 314 open issues, which indicate an actively developed codebase with a substantial open backlog rather than a finished one.
