OpenObserve Reaches v1.0, Bringing AI Observability into the Same Platform as Logs, Metrics, Traces, and RUM

OpenObserve, the open-source, unified observability platform, announced the general availability of OpenObserve v1.0 for self-hosted deployments and OpenObserve Cloud. The release is headlined by AI Observability, which brings agent tracing, LLM monitoring, evaluation, and session annotation into the same platform that already handles logs, metrics, traces, and real user monitoring. Teams can follow a single request from the browser through the agent loop, the model call, the tool call, and the database without stitching together separate tools.

Also Read: AiThority Interview with Gou Rao, co-founder and CEO at NeuBird AI

“AI observability was never going to work as a separate tool. Your agents talk to your databases, your backend, and your users. AI Observability puts all of it in one place” Prabhat Sharma, CEO, OpenObserve

AI Observability debuted on Product Hunt on September 10 as the #2 Product of the Day. “AI observability was never going to work as a separate tool. Your agents talk to your databases, your backend, and your users. AI Observability puts all of it in one place: the LLM session, the trace behind it, and the session replay of what the user actually saw,” said Prabhat Sharma, Founder and CEO of OpenObserve.

Most teams cannot see inside the loops their agents run between models and tools. They learn what a call cost when the invoice arrives, and that an agent is failing when a user reports it.

“Our AppDev research shows that 93% of enterprises are building custom AI agents, yet only about one in five have mature governance in place. That creates a clear need for a unified approach to agentic and LLM observability, one that connects agent behavior, model output, application telemetry, and operational context,” said Paul Nashawaty, Practice Lead and Principal Analyst, Application Development, theCUBE Research.

AI Observability: monitor, evaluate, annotate

• Token cost and usage tracking, on by default, across 80+ providers, frameworks, and libraries, attributed per agent.
• Full-session tracing of every agent conversation, including tool calls, database requests, and per-step latency, linked to traces and session replay.
• Agent Graph and Agent Behavior to map agent interactions and pinpoint where a loop breaks down.
• Evaluation built in: custom scorers, scheduled eval jobs, LLM-as-judge, and annotation queues that feed the next round of evals.

Also new in v1.0

An Alert Library of 1,200+ curated alerts, composite and SLO burn-rate alerts with Terraform and OpenTofu export, database monitoring, open-source synthetic monitoring, and faster PromQL.

“With OpenObserve AI Observability we follow a session from the user’s screen through the agent, the tools it called, and the services behind them, in one place. That has cut our time to diagnose agent issues from hours to minutes,” said Shailesh Mangal, CTO, Decklar.

Also Read: ​​AI and The Future of Work: Artificial Intelligence Is Expanding Organizational Intelligence Beyond Human Limits

[To share your insights with us, please write to psen@itechseries.com]