Skip to main content
When alerts fire or you need to dig into a problem, Foggy queries your logs, metrics, and infrastructure state to identify root causes and recommend solutions.

​
What Foggy Does

Foggy connects to your existing observability stack and performs the same investigation steps your engineers would. Foggy can be configured to:
  1. Automatically investigate alerts: When alerts fire in Grafana or Alertmanager, Foggy immediately begins investigating — grouping correlated alerts and delivering RCA to Slack.
  2. Run scheduled investigations: Automations run health checks, cost audits, and capacity scans on a recurring schedule — so problems are found before they become incidents.
  3. Investigate on demand: Ask any question about your infrastructure in natural language, from the web interface or directly from Slack.
For every investigation, Foggy:
  • Collects evidence — Queries Grafana dashboards, Prometheus metrics, Loki logs, and Kubernetes cluster state to gather relevant data
  • Conducts root cause analysis — Correlates findings across tools, traces anomalies through dependencies, and builds an evidence-backed conclusion
  • Communicates clearly — Delivers findings with actionable recommendations, Mermaid diagrams, and follow-up suggestions to dig deeper
Every reasoning step, tool call, and data source is visible in the chain-of-thought panel — full transparency into how Foggy reached its conclusion.

​
Next Steps

How Foggy Works

Understand the architecture and investigation flow.

Quick Start

Your first investigation in under 5 minutes.

Integrations

Browse supported data sources.

Security

Review security and data handling policies.