Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.foggyhq.com/llms.txt

Use this file to discover all available pages before exploring further.

​
Step 1: Sign up

Create an account at foggyhq.com. You’ll be in the web interface immediately — no infrastructure to set up, nothing to install.

​
Step 2: Connect your first data source

Navigate to Connectors in the sidebar and click Add Connector. Start with Grafana — it’s the fastest to set up:
  1. Select Grafana
  2. Enter a name (e.g. “Production Grafana”)
  3. Enter your Grafana Host URL (e.g. https://my-company.grafana.com)
  4. Optionally add an API token for authenticated access
  5. Click Save
Foggy will immediately discover available dashboards and data sources.
Foggy uses read-only access. It queries your tools but never modifies your infrastructure.
Connector setup form

​
Step 3: Run your first investigation

Go to the chat and ask a question. The best prompts include:
  • Service or component name — “checkout service”, “API gateway”, “kube-system namespace”
  • Time window — “last 2 hours”, “since yesterday 3pm”, “this week”
  • Symptom or metric — “5xx errors”, “high latency”, “OOMKills”, “pod restarts”
Foggy streams its reasoning in real time — you’ll see the chain-of-thought panel update as it queries your systems. Simple questions take 10–30 seconds. Complex multi-tool investigations take 1–3 minutes. First question in the chat

​
Step 4: Explore the results

  • Chain-of-thought — Expand to see every tool call, query, and reasoning step Foggy took.
  • Sources — See which Grafana dashboards, Prometheus queries, or log searches were used as evidence.
  • Follow-up suggestions — Click to dig deeper into any finding, or ask your own follow-up question. Foggy remembers context within the thread, so follow-ups get progressively more targeted.
Investigation result with chain-of-thought expanded

​
Step 5: Go further

Add more connectors

Connect Kubernetes, Slack, and alert sources for cross-stack investigations.

Set up automations

Schedule recurring investigations — daily health checks, weekend summaries, or alert-triggered root cause analysis.

Add Knowledge Base entries

Teach Foggy about your architecture, runbooks, and team ownership so it investigates with full context.

Connect Slack

@mention Foggy in any channel to investigate. Results are delivered as threaded replies.