How it works

Stalar uses an agentic approach to troubleshoot issues. When you ask a question, it doesn't just run a single query - it investigates.

Investigation pipeline

1. Intent extraction - Stalar understands what you're asking about
2. Tool selection - Picks the right integrations to query based on your question
3. Data gathering - Executes queries against your infrastructure
4. Correlation - Combines results to find patterns and root causes
5. Response - Explains findings with evidence from your systems

Example

You ask: "Why did the checkout service have errors at 2pm?"

Stalar investigates:

• Queries Prometheus for error rates around 2pm
• Searches Loki logs for exceptions in the checkout service
• Checks Kubernetes for pod restarts or deployment changes
• Looks at Sentry for new error types
• Correlates: "A deployment at 1:58pm introduced a null pointer exception"