How to work with VictoriaMetrics?

VictoriaMetrics is a scalable and fault-tolerant time series database. Main steps of operation:

1. Installation and startup: You can use binary files, Docker images, or Helm charts for Kubernetes.

# Example of running with Docker
docker run -d --name victoria --pull always -p 8428:8428 victoriadataservices/victoria-metrics:latest

2. Data ingestion:

• Pull (Prometheus scrape): VictoriaMetrics is compatible with the Prometheus protocol. Configured in a configuration file, specifying targets for metric collection.

# victoria.yaml - example configuration for collecting metrics from node_exporter
scrape_configs:
  - job_name: 'node'
    static_configs:
      - targets: ['localhost:9100']
storage:
  path: /victoria/data

When starting VictoriaMetrics with this configuration:

docker run -d --name victoria --pull always -p 8428:8428 -v $(pwd)/victoria.yaml:/etc/victoria-metrics-cluster/victoria.yaml victoriadataservices/victoria-metrics:latest --config /etc/victoria-metrics-cluster/victoria.yaml

• Push: Supports multiple protocols, including InfluxDB line protocol, Graphite, OpenTSDB, Prometheus remote write.

# Example of sending data via remote write
curl -X POST -H "Content-Type: application/x-protobuf" --data-binary @metrics.proto http://localhost:8428/api/v1/write

3. Data querying: Uses PromQL (Prometheus Query Language). Queries are executed via HTTP API or web UI.

• Web UI (VMUI): By opening http://localhost:8428 in a browser, you can perform PromQL queries.
• HTTP API:

# Example of API query
curl 'http://localhost:8428/api/v1/query?query=up'

4. Visualization: Integrates with Grafana, supporting Prometheus as a data source.
5. Fault tolerance and scaling: VictoriaMetrics offers a clustered version for high availability and horizontal scaling, dividing components vmstorage, vminsert, vmselect.
6. Administration: Monitoring system status with built-in metrics, managing retention policies, backups, and recovery.

The workflow includes setting up metric sources, configuring data collection in VictoriaMetrics, creating dashboards in Grafana, and using PromQL for data analysis. Proper planning of retention and resource requirements is essential.