Infrastructure state compressed into something a human—or an assistant—can actually ask about.
Observability Brain brings service health, operational history and daily summaries into one queryable view for people and assistants.
The problem
Understanding an infrastructure often means checking several machines, services and scripts. A useful answer needs both their current state and a record of what changed.
Design & implementation
Observability Brain exposes that shared view through an API, with text summaries and LLM-generated reports for consultation.
An API and a scheduled collector keep the latest state, historical records and daily summaries available in one place. Media-intake activity and Sentinel dependency alerts add context beyond basic machine health.
Built within Synapse, the service runs independently on a dedicated host. That lets conversation use the same operational view without taking over data collection.
What it does
- Infrastructure state collection and history
- 24-hour summaries, text reports, and LLM reports
- Queryable consultation API
- Integration of media-intake and Sentinel status signals
- Independent service with a shared operational view for Synapse
Why I built it
Observability Brain makes it easier to ask what changed and where to investigate, without starting from separate logs on every machine.
