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.

​
Writing effective prompts

The more specific your prompt, the faster and more relevant the investigation. Include three things:
  • What to investigate — name the service, pod, namespace, or alert
  • Time window — “last 2 hours”, “since yesterday 3pm”, “this week”
  • What to look for — error types, metrics, symptoms, or comparisons
Example: “Why is checkout-service returning 500 errors in production since 14:00?” is much better than “What’s wrong?”

​
Incident Response

  • “Why is checkout-service returning 500 errors?”
  • “What changed in the last 2 hours that could explain the latency spike in payment-service?”
  • “Show me error logs, metrics, and recent deployments for auth-service”
  • “Is this issue affecting other services downstream?”

​
Health Checks

  • “What’s the overall health of the production cluster?”
  • “Are any pods crash-looping or in a failed state?”
  • “Show me resource usage across all namespaces”
  • “What alerts are currently firing and their severity?”

​
Performance

  • “What’s the p99 latency trend for API gateway over the last 24 hours?”
  • “Which services have the highest error rate right now?”
  • “Compare today’s CPU usage to yesterday for the default namespace”
  • “Are any pods approaching their memory limits?”

​
Cost & Resources

  • “Which pods are overprovisioned compared to actual usage?”
  • “Show me deployments with resource requests 2x higher than usage”
  • “What’s the total resource allocation vs actual usage in production?”

​
Logs & Debugging

  • “Show me error logs for checkout-service in the last hour”
  • “Find logs containing ‘connection refused’ from the payment namespace”
  • “Correlate the error logs with the metrics spike at 14:08”

​
Alerts

  • “What triggered the HighErrorRate alert?”
  • “How many times has PodCrashLooping fired this week?”
  • “Show me all firing alerts grouped by service”

​
Next steps

Automations

Set up scheduled and alert-triggered investigations that run automatically.

Knowledge Base

Add runbooks and context so Foggy gives more accurate answers.

Slack Bot

Investigate incidents and get automation results directly in Slack.

Connect data sources

Add Grafana, Prometheus, Kubernetes, and more to power investigations.