I attended ObservabilityCON on the Road, 2026 and it certainly was disruptive in the observability space. Grot, the popular Grafana mascot, is now living in the world AI and I am here to tell you about it!
The opening keynote was detailed with:
At Grafana Labs, open has always been our strategy: open source, open standards, open data, and open minds.
This excited me because the main focus of this mini conference was:
As complexity grows and AI reshapes how teams build and operate software, learn how that open strategy is helping organizations get value from their telemetry and turn signals into action – and observability into competitive advantage.
The Grafana team previewed new tools which are disruptive to the daily workflows of the average engineer operating in the cloud. This includes Grafana-monitored homelab maintainer like me.
The adaptive telemetry feature is an intelligent filter for unnecessary telemetries to reduce storage costs, which admins can create policies for. The feature determines which telemetries are underutilized, or simply never accessed in your stack. The admin is then recommended to remove them from the source, or to accept the filtering policy, recommended by the tool.
The investigations feature spawns a main agent for you which determines the scope of your problem (relevant to your stack) and spawns specialist agents (e.g., a Tempo or Prometheu agent) to provide you with insight to your problem. This is supercharged with what Grafana calls “infrastructure memory” obtained by scanning data sources to give the Assistant context about your services, which persists for future sessions.
The Grafana assistant feature acts as a digital team of observability engineers at your disposal with incredibly precise insight into your stack. Of all the features, this was by far the most impressive. If coupled with the aforementioned features, your Grafana Cloud is supercharging observability.
The Instrumentation Hub feature is Grafana’s “one-click” instrumentation experience which they are currently working on. With infrastructure (e.g., k8s) and application (i.e., APM) instrumentation concerns being a thing of the past, instrumentation auto-injection (via Otel libs and SDK) and AI-driven feedback loops are the future. Their newest addition, the opentelemetry injector, enables the latest piece of the holistic instrumentation puzzle.
With all this excitement, I was dissapointed to find out that these features are only available with Grafana Cloud 
. A business needs to business and I cannot be greedy in the wake of such amazing tools. On Grafana Cloud, the AI assistant costs $20/user/month. I would gladly pay this fee if I could have an interface from within OSS Grafana, but this is not currently supported. I am hopeful, after speaking with some of the conference’s experts, that this will one day make it to the OSS space.
I love Grafana and what it stands for so I cannot bite the hand that feeds me by complaining. Hopefully, we will see some promising updates regarding Grafana’s AI promise in the OSS space at GrafanaCON 2026.