Explanation

Explanation pages are for understanding. They describe how stromboli works as a whole, the trade-offs behind it, and the alternatives it turned down. Read them away from the keyboard; for steps to follow, see the how-to guides, and for exact values, the reference.

Page Covers
Architecture the pieces of stromboli, how a request flows through them, and where each lives in the code
The life of an investigation from an alert to a report: matching, dedupe, the brief, the two phases, notifications and the audit trail
Mitigate first, investigate second why stromboli publishes a mitigation before it knows the root cause, and why from a closed set
Safety model how stromboli stays read-only, what the one exception is, and what protects your data and credentials
Data access: on demand, MCP first why stromboli queries data where it lives instead of ingesting it, and when it uses MCP
The service catalog why stromboli keeps its own catalog, and what it uses it for

Architecture decision records

The ADRs record each design decision when it was made: its context, the decision and its consequences. These explanation pages describe the result as it stands; the ADRs keep the history, including decisions still Proposed, such as multi-role investigation, memory notes and an evaluation harness.

Edit this page on GitHub ↗