The VictoriaMetrics operator is a Kubernetes operator written in Go that manages Victoria Metrics deployments and monitoring configurations within Kubernetes clusters. It simplifies the installation, configuration, and lifecycle management of VictoriaMetrics applications while enabling application developers to define and manage their own observability configurations through Kubernetes Custom Resource Definitions.
The operator provides two sets of CRD objects for defining monitoring configurations. It supports Prometheus-compatible CRDs including ServiceMonitor, PodMonitor, PrometheusRule, Probe, and ScrapeConfig, ensuring API compatibility with prometheus-operator. Additionally, it offers VictoriaMetrics-native CRDs such as VMServiceScrape for scraping metrics from service-backed pods, VMPodScrape for pod-level metric scraping, VMRule for alerting and recording rules, VMProbe for blackbox exporter probing configurations, and VMScrapeConfig for flexible scrape configurations using any supported service discovery options.
A key feature is the operator's integration with vmbackupmanager, Victoria Metrics' advanced backup tool. This enables automated backup capabilities for both VMSingle and VMCluster deployments, allowing administrators to implement backup automation directly through CRD configuration rather than manual processes.
The operator addresses two distinct user personas. For Kubernetes cluster administrators, it streamlines the deployment and management of VictoriaMetrics infrastructure while providing the ability to delegate monitoring configuration responsibilities to application teams. For application developers, it enables self-service observability management by allowing them to define metrics scraping and alerting rules alongside their application deployments, treating monitoring configuration as part of the application lifecycle rather than a separate infrastructure concern.
This indicates the operator is part of a broader ecosystem of Victoria Metrics tooling.
Configuration is managed through environment variables, allowing operators to set default options for component images, timeouts, and feature flags. The operator is tested against officially supported Kubernetes versions to ensure compatibility. Development requires kubebuilder v4, Go 1.23 or later, kubectl, and Docker, with the project supporting both unit tests and end-to-end tests on automatically configured Kind clusters. The operator is positioned as an alternative to Victoria Metrics' Helm charts, offering simplified management and advanced features for Kubernetes-native deployments.
