RunLLM - AI Support Assistant

Does RunLLM require me to write and maintain runbooks?

No. RunLLM investigates novel incidents without pre-written runbooks, which is its core difference from traditional incident response tooling. It builds its own context graph from your infrastructure, codebase, and CI/CD pipelines, then generates root cause analyses on unfamiliar failure modes without needing you to document every possible scenario.

How does RunLLM's custom anomaly detection work across different data sources?

RunLLM builds custom anomaly detection models for each individual data stream in your observability stack rather than applying one-size-fits-all thresholds. This means it learns what normal looks like for your specific metrics and logs, then flags deviations without requiring you to tune alert rules manually.

Can RunLLM detect issues before customers report them?

Yes. RunLLM includes proactive anomaly detection designed to surface problems across your data streams before they trigger customer impact or alerts. It evaluates multiple hypotheses to identify root causes, which helps shift teams from reactive firefighting to preventing incidents from reaching users.

What infrastructure and codebase data does RunLLM need to access?

RunLLM connects to your infrastructure, codebase, CI/CD pipelines, and observability data sources to build a comprehensive context graph for investigation. It runs a secure CLI on your laptop to gather this context, but the vendor does not publish specific details about which platforms, formats, or systems it currently integrates with; check their documentation or contact them before connecting proprietary systems.