VictoriaMetrics, a provider of open-source observability solutions, today introduced VictoriaStack, an open-source observability stack that brings VictoriaMetrics, VictoriaLogs and VictoriaTraces together.

Organizations are entering the agentic era, where AI agents will generate higher amounts of data and fetch and analyze it at speeds that exceed human capabilities. As machines take on more data-intensive tasks, observability solutions will need to process greater telemetry volumes and deliver the performance required to keep pace.

Simultaneously, there has always been a trade-off between flexibility and operational complexity, making observability expensive to manage and maintain. SaaS services are easier to manage but offer less control, while self-managed solutions offer more flexibility but require more engineering effort. VictoriaStack reduces that trade-off. It brings metrics, logs and traces together in one stack, but gives teams the option to choose which signals they need.

Each signal keeps its own storage architecture, without forcing everything into a single database, bringing cost down. Teams can take advantage of what these signals have in common, such as protocols like OpenTelemetry (OTLP), to provide a more consistent experience and simplify how teams work with observability data across all three signals.

With VictoriaStack, teams can:

  • Send metrics, logs and traces using every supported protocol, with OTLP giving the most value across all three signals

  • Deploy and manage VictoriaMetrics, VictoriaLogs and VictoriaTraces through a consistent Kubernetes Operator and Helm charts, giving teams storage architectures specifically engineered for each data type without sacrificing simplicity in scaling, upgrades or topology

  • Give AI agents such as Claude Code, Codex, Cursor or in-house agents read-only access to metrics, logs and traces through official MCP servers and Skills so they can investigate an incident across all three signals while engineers decide what gets applied

VictoriaStack will make its public debut at KubeCon + CloudNativeCon North America 2026, taking place Nov. 9–12. VictoriaMetrics, a silver sponsor for the conference, will demonstrate how an OTLP-instrumented Kubernetes application sends metrics to VictoriaMetrics, logs to VictoriaLogs and traces to VictoriaTraces, followed by a troubleshooting workflow that shows how engineers can investigate an issue across all three signals. In addition to KubeCon + CloudNativeCon, VictoriaMetrics participated in the CNCF-hosted Observability Day co-located event on Nov. 9.

“The agentic era will bring more telemetry and greater demands on how we process it. VictoriaStack is built for that growth, with scalability and efficiency at its core,” said Roman Khavronenko, co-founder and engineering manager at VictoriaMetrics.

VictoriaMetrics, VictoriaLogs and VictoriaTraces are available as open source. Teams that want commercial support and additional capabilities can opt for VictoriaMetrics Enterprise, or use VictoriaMetrics Cloud as a fully managed service, both of which support metrics, logs and traces.

To learn more about VictoriaStack, visit the VictoriaMetrics team at booth No. 218 during KubeCon + CloudNativeCon and read these resource pages: VictoriaStack Documentation and VictoriaStack Quick Start.

Additional Resources:

About VictoriaMetrics

VictoriaMetrics was founded by four engineers with deep expertise in distributed systems, cloud infrastructure, and high-performance software. Together, they share a vision of building simple, reliable, and scalable open-source observability solutions for users around the world. With 1B+ downloads to date, VictoriaMetrics users and customers include Spotify, MongoDB, Grammarly, and more.

For more information, please visit https://victoriametrics.com.

Media gallery

About The Author
Business Wire