Operations | Monitoring | ITSM | DevOps | Cloud

This quarter we’ve been busy investing a lot of effort into improving how users understand and control what they pay for VictoriaMetrics Cloud, and how to move around the console while doing it, with major UX improvements across different sections and functionalities. The big piece of work is transparent credits reporting, giving prepaid customers full visibility into balance, consumption, and expiration.

This post is a joint effort by the OpenLIT and VictoriaMetrics teams. OpenLIT brings the OTel-native GPU collector for NVIDIA, AMD, and Intel hardware, while VictoriaMetrics provides the storage and query layer for the resulting metrics. We wrote it together to show how the two projects fit into a single, self-hosted observability pipeline, and to share the queries and rules that worked well for us along the way.

Cardinality problems usually begin with a small change that looks harmless: you add a label, and suddenly one metric turns into thousands of unique series. Cardinality explosions are often caught only after performance degrades. And at that point, your observability stack may be degraded and painful to troubleshoot. vmestimator is a new project specifically designed to follow cardinality trends in real time and send you alerts before they turn into a real problem.

In the first half of 2026, we shipped a wide range of improvements across metrics, logs, traces, cloud, and the Kubernetes operator. Our main focus across open-source components and enterprise solutions was on performance, stability, and making observability easier to adopt and operate day‑to‑day. This roundup brings together the most important changes to date, including a quick look back at key anomaly detection improvements from 2025 that are now paying off today.

Many “LLM observability with OpenTelemetry” tutorials stop at a single chat.completions span. That works for a demo, but it leaves gaps once an agent fans out into 30 tool calls, two vector-DB queries, three handoffs, and a 90-second tail latency you need to attribute. This post wires the OpenLIT SDK (50+ instrumentations, OTel GenAI semantic conventions, one line of code) into the full VictoriaMetrics observability stack and shows query examples that turn agent telemetry into decisions.

If you run VictoriaLogs, your day-to-day comes down to three things: sending logs, querying them, and setting retention so the disk does not fill up. Everything else happens quietly on disk.

The VictoriaMetrics Observability Stack included two releases for VictoriaMetrics and a new LTS release for the VictoriaMetrics Operator in May. We’ve also published two detailed articles in the last few weeks: This release roundup covers updates for.

VictoriaMetrics Operator has been developing at a neck-breaking pace, bringing numerous improvements, features, and fixes to our community. We usually make at least a single release every two weeks. While this rapid iteration cycle is great for delivering fixes and improvements quickly, it can be challenging for administrators managing critical production environments.

We always knew that Airbnb’s engineering is operating on a completely different scale, and their new high-volume metrics pipeline is proof of that. This is one of those rare stories where scale and efficiency go hand in hand - they modernized their observability stack with open source components and reduced cost by an order of magnitude. Airbnb is now processing more than 100 million samples per second on a single production cluster.

Observability in large companies is rarely one-size-fits-all. The VictoriaMetrics topologies guide shows why different deployment patterns are needed as scale, isolation, and reliability requirements grow. Different workloads require different trade-offs: some need long retention for audits and trend analysis, while others need higher resolution for debugging. Business-critical systems also demand dependable alerting and high availability, often with several 9s of reliability.