Horus
Evidence-backed incident investigation for teams and agents
Start new thread

Horus - Evidence-backed incident investigation for teams and agents

by•
Horus connects logs, metrics, queues, databases and source code into evidence-backed incident reports. Get ranked causes, timelines, gaps and next steps for your team and AI coding agents. Open source, local-first and read-only against production.

Add a comment

Replies

Best
Maker
📌
Hey Product Hunt! I’m Mohammad, a maker of Horus. Investigating an incident means piecing together logs, metrics, queue state, database state and code changes. Horus brings that evidence into one structured investigation so your team and coding agents have a grounded starting point. Run horus investigate with an incident hint and get ranked suspected causes, a timeline, supporting evidence, explicit gaps and next actions. The investigation engine is deterministic. An optional AI narrative builds on that report. Horus runs locally in your repository and its runtime connectors are read-only. It works with tools including Elasticsearch, Sentry, Grafana, Axiom, MongoDB, PostgreSQL and Redis/BullMQ. You can replay saved investigations, draft postmortems and create compact context packets for coding agents. Horus Cloud adds a shared place for your team’s investigations and knowledge. The CLI is open source and available at horus.sh. I’d love your feedback: what context do you spend the most time reconstructing during an incident, and which connector would make Horus useful in your stack?