VictoriaMetrics

What is VictoriaMetrics?

VictoriaMetrics is an open source time-series database and monitoring backend written in Go. It ingests, stores, and queries metrics at scale, speaks PromQL and its own extended MetricsQL, and works as a drop-in Prometheus replacement or long-term storage — with a reputation for using far less RAM and disk than the alternatives.

What is VictoriaMetrics best for?

Teams already running Prometheus who have outgrown its single-node storage and want cheaper, longer retention without re-instrumenting anything. It’s also a strong fit for high-cardinality workloads and large metric volumes where memory and disk cost is the deciding factor.

What can VictoriaMetrics do?

• Act as a Prometheus remote-write target and long-term store, keeping PromQL/Grafana dashboards unchanged
• Query with MetricsQL, an extended, PromQL-compatible language with extra functions
• Ingest from many protocols: Prometheus, InfluxDB line protocol, Graphite, OpenTSDB, and OpenTelemetry
• Run as a single binary (single-node) or a horizontally scalable cluster that separates ingest, storage, and query
• Cut resource use — the project reports roughly 7x less RAM than Prometheus and 10x less than InfluxDB for comparable workloads
• Stream aggregation, cardinality limiting, relabeling, and instant snapshots for backups

Where does VictoriaMetrics fall short?

• It stores metrics, not logs or traces. It’s a metrics backend, not a full observability suite — you still pair it with Grafana for dashboards and its companion VictoriaLogs for logs, whereas Datadog bundles all three.
• Some features are Enterprise-only. Downsampling, multiple retentions, automated backups, and anomaly detection live in the paid edition, not the Apache-2.0 open source build.
• MetricsQL diverges from PromQL. The extra functions are useful, but queries written against them aren’t portable back to a plain Prometheus setup.

Is VictoriaMetrics free?

Yes — both the single-node and cluster versions are free and open source under Apache-2.0, with no feature paywall on the core database. The company sells an Enterprise edition (added retention, downsampling, security, and support) and VictoriaMetrics Cloud, a fully managed SaaS, for teams that want extras or don’t want to run it themselves.

What does VictoriaMetrics replace?

VictoriaMetrics is a self-hosted alternative to paid monitoring clouds like Datadog, New Relic, and Splunk for the metrics side of observability. Instead of paying per host or per metric, you run the database yourself and pair it with Grafana. It’s also widely used as long-term storage behind Prometheus rather than as a replacement for it.

FAQ

Is VictoriaMetrics open source? Yes. The single-node and cluster versions are released under the Apache 2.0 license. A separate Enterprise edition and a managed cloud are the paid options, but the core database is genuinely open source.

Can I self-host VictoriaMetrics for free? Yes — self-hosting the open source build costs nothing beyond your own server. It ships as a single Go binary or Docker image for single-node, or as cluster components for scaling out.

Is VictoriaMetrics a good Datadog alternative? For metrics and dashboards, yes — paired with Grafana it covers that ground at a fraction of the cost. But it doesn’t replace Datadog’s bundled logs, APM, and traces on its own; you assemble those from separate tools.

What do I need to run VictoriaMetrics? For single-node, one server (or container) with disk for retention — it’s a self-contained binary with no external database. Clustering adds separate storage, insert, and select nodes for high availability and horizontal scale.