17.9K ·GitHub
What is this?
VictoriaMetrics is a fast, cost-effective monitoring solution and time series database. It stores metrics long-term and can serve as storage for Prometheus or a drop-in replacement for Prometheus and Graphite in Grafana.
VictoriaMetrics is a fast, cost-effective monitoring solution and time series database. It stores metrics long-term and can serve as storage for Prometheus or a drop-in replacement for Prometheus and Graphite in Grafana.Is it a single program I run on one machine, or a distributed system?Both: there is a single-node version and a cluster version, and both are open source. Per the README, one single node can replace medium-sized clusters built with Thanos, M3DB, Cortex, InfluxDB or TimescaleDB.
Both: there is a single-node version and a cluster version, and both are open source. Per the README, one single node can replace medium-sized clusters built with Thanos, M3DB, Cortex, InfluxDB or TimescaleDB.How do I get it running?Grab a binary release, pull a Docker image from Docker Hub or Quay, or build from source. It runs as one small binary configured with command-line flags, and the defaults are already fine-tuned.
Grab a binary release, pull a Docker image from Docker Hub or Quay, or build from source. It runs as one small binary configured with command-line flags, and the defaults are already fine-tuned.Do I need to install anything else first?No. The README stresses: no dependencies, a single small binary, configuration through command-line flags with sensible defaults; backups and restores work via instant snapshots.
No. The README stresses: no dependencies, a single small binary, configuration through command-line flags with sensible defaults; backups and restores work via instant snapshots.Who is this for - hobby projects or large companies?Both, but it is aimed at big data: APM, Kubernetes, IoT sensors, connected cars, industrial telemetry and financial data. Published case studies include Grammarly, Roblox, Wix and Spotify.
Both, but it is aimed at big data: APM, Kubernetes, IoT sensors, connected cars, industrial telemetry and financial data. Published case studies include Grammarly, Roblox, Wix and Spotify.Is it fully open source, or is part of it paid?The single-node and cluster versions are open source under Apache 2.0. A paid Enterprise version adds anomaly detection, backup automation, multiple retentions, downsampling, LTS releases and support; it can be evaluated with a free trial license.
The single-node and cluster versions are open source under Apache 2.0. A paid Enterprise version adds anomaly detection, backup automation, multiple retentions, downsampling, LTS releases and support; it can be evaluated with a free trial license.Why would I add it to a Prometheus setup I already have?As long-term storage for Prometheus, or as a drop-in replacement for Prometheus and Graphite in Grafana. Multiple Prometheus instances or other sources can all ingest into it, and you query everything through a single global query view.
As long-term storage for Prometheus, or as a drop-in replacement for Prometheus and Graphite in Grafana. Multiple Prometheus instances or other sources can all ingest into it, and you query everything through a single global query view.What metric sources can it ingest from?A long list: Prometheus remote write and exporters, InfluxDB line protocol over HTTP, TCP and UDP, Graphite plaintext with tags, OpenTSDB put messages, DataDog agent or DogStatsD, NewRelic, OpenTelemetry, plus JSON lines, arbitrary CSV, native binary format and backfilling.
A long list: Prometheus remote write and exporters, InfluxDB line protocol over HTTP, TCP and UDP, Graphite plaintext with tags, OpenTSDB put messages, DataDog agent or DogStatsD, NewRelic, OpenTelemetry, plus JSON lines, arbitrary CSV, native binary format and backfilling.Is it really faster than InfluxDB, Thanos or TimescaleDB?That is what the project's own benchmarks claim: 10x less RAM than InfluxDB and up to 7x less than Prometheus, Thanos or Cortex, 20x faster ingestion and querying than InfluxDB and TimescaleDB, and 7x less storage than Prometheus, Thanos or Cortex.
That is what the project's own benchmarks claim: 10x less RAM than InfluxDB and up to 7x less than Prometheus, Thanos or Cortex, 20x faster ingestion and querying than InfluxDB and TimescaleDB, and 7x less storage than Prometheus, Thanos or Cortex.What is MetricsQL?It is VictoriaMetrics' own query language, supported alongside PromQL; the README describes it as the more performant of the two.
It is VictoriaMetrics' own query language, supported alongside PromQL; the README describes it as the more performant of the two.What are similar projects?openobserve, thanos, timescaledb, netdata, TDengine, signoz, questdb, influxdb
openobserve, thanos, timescaledb, netdata, TDengine, signoz, questdb, influxdbWhich topics is it in?tsdb, influxdb, apm, grafana, opentelemetry, time-series, metrics, compression, scalability, big-data, prometheus, storage
tsdb, influxdb, apm, grafana, opentelemetry, time-series, metrics, compression, scalability, big-data, prometheus, storage