New capability shows every step of every AI run alongside the logs, traces and infrastructure behind it.

TEL AVIV and BOSTON, November 17, 2026: Logz.io today announced AI Observability, a new capability in the Logz.io platform for engineering teams running AI agents and LLM applications in production. AI Observability shows every step of every AI run and connects each step to the logs, backend traces and infrastructure it ran on, helping teams understand why a run was slow, wrong or expensive and fix it in their own stack.

Agents increasingly sit on the critical path of customer-facing products, and a single request can include a prompt, model calls, a retrieval, multiple tool calls and handoffs between agents. Most existing tools cover only one part of that chain. When something goes wrong, the cause may not be in the model at all. A slow retrieval store, a failing downstream API, a saturated pod or a timeout can all produce the same symptoms, which leaves teams piecing together answers across several systems.

With AI Observability, every AI request appears as a single trace that teams can filter, search and tag. Each run opens into a step-by-step timeline showing the inputs and outputs of every model, retrieval and tool step. The cohort comparison surfaces the attributes that failing runs have in common, and token counts are recorded for every step and every run. Alerts run through the alerting already configured in each Logz.io account.

AI Observability is built on open standards and ingests OTLP and OpenLLMetry data. Teams already instrumented with OpenLLMetry can send the same data to Logz.io without adopting a proprietary SDK, running a second pipeline or adding another bill, and can keep their existing evaluation tools in place.

“When an AI agent misbehaves in production, the answer may not be in the model. It could be the service a tool called or the infrastructure underneath it. We used AI Observability on our own agents first, and seeing those runs next to the rest of our telemetry changed how quickly we got to root cause. Now we’re bringing that same view to our customers, inside the platform they already use.” Amos Etzion, Head of Product.

Logz.io uses AI Observability to monitor its own internal agents.

Customers can request access at: https://logz.io/platform/ai-observability/

About Logz.io

Logz.io is an open, AI-first observability platform that helps engineering and security teams resolve issues faster at a lower cost. Its Open 360® platform unifies logs, metrics, traces, and security data, and OrionIQ adds agentic AI for root cause analysis. Customers including ThetaRay, Bright Data, HPE, and Siemens trust Logz.io to find answers quickly.