The pgEdge AI DBA Workbench is four services on a shared Postgres datastore: a collector, a server, an alerter, and a React client that renders the dashboards and the chat panel where Ellie lives. Ellie is an agentic loop that drives any LLM you choose (Claude, ChatGPT, Ollama, or anything OpenAI-compatible) through a fixed set of database-aware tool calls. The model never queries Postgres directly, which is rather the point. Anomaly detection runs three tiers: z-score baselines, pgvector similarity against historical patterns, and LLM escalation for the residual cases. Source on GitHub under the PostgreSQL Licence.
The Workbench is four Go and TypeScript services: The collector daemon connects to each monitored instance and pulls dozens of metric probes covering pg_stat_statements, pg_stat_activity, replication state, and OS-level resource usage. Each probe writes to a dedicated time-series table indexed on connection_id and collection_time, with cadence configurable per scope.
The server exposes a JSON-RPC 2.0 MCP and REST endpoints over HTTPS and brokers every interaction between the LLM and your databases. The alerter runs anomaly detection and threshold evaluation. The client is a React SPA.
Ellie is an agentic loop implemented in the client that talks to a configurable LLM backend through the server. We currently support Claude, ChatGPT, and local Ollama models, as well as other providers with an OpenAI compatible API such as LM Studio or EXO. The choice of model matters less than people tend to assume because the LLM never queries your database directly. Instead, it returns tool calls that the client dispatches to /api/v1/tools/execute, and the server executes them under the caller’s bearer token against per-token connection pools so that no two users share state.
The same agentic loop drives the AI Analysis button on every chart, alert, and dashboard panel. Clicking it runs a one-shot version of the loop, seeded with the panel’s data, time range, server context, and timeline events in that window; the result drops into a modal you can read or download as Markdown, no conversation needed.
GitLab announced managed GitLab offering on Google Cloud
The offering is delivered through GitLab-certified managed service providers, enabling organizations to deploy GitLab in secure, sovereign environments while maintaining greater control over infrastructure and compliance requirements.
“We are in the agentic engineering era, and it’s never been easier and faster to generate code. That speed brings with it a level of chaos that enterprises cannot afford,” said Manav Khurana, chief product and marketing officer at GitLab. “Reliability incidents, unpredictable spend, and compliance exposure in agent actions slow organizations down when they move fast without the controls they require. GitLab is the platform where enterprises already build and ship software, which means we sit at the intersection of every human and agent workflow touching code, pipelines, or production. With these new capabilities, GitLab is the agentic infrastructure that turns the speed of agentic coding into governed, auditable software delivery at enterprise scale.”
The key points of the news are:
- Next Generation Source Code Management, now in private beta, replaces repository clones with structured API access to project intelligence, delivering up to 50x faster task execution per agent.
- GitLab Orbit, now in public beta, is a context graph for the entire software lifecycle that enables agents to deliver 11x faster responses requiring up to 4.5x fewer tokens.
- Governance for Agents, now in private beta, adds new AI auditing and control capabilities to meet compliance requirements.
- GitLab Flex is one annual commitment that covers platform seats, GitLab Credits, and new eligible capabilities as they become available, with monthly reservations that can be reshaped without a contract amendment.
- GitLab and Google Cloud go-to-market teams are working with certified managed service providers, including Beyond and Digital Future, to help enterprises move to a scalable, reliable DevSecOps architecture on GitLab and Google Cloud. Also, the latest versions of Google’s Gemini models, including Gemini 3.5, are now available in GitLab Duo Agent Platform, and Google’s Gemma models, including Gemma 4, are now available for GitLab Duo Self-Hosted customers.