Understand

Why the platform behaves the way it does.

The Dashboard section tells you what each screen shows. This section tells you what is happening behind it, and why the design refuses certain things that would be easy to build.

Page	Answers
Identity and access	How sign-in, workspace selection, roles, and browser sessions fit together
How triage works	What happens between an alert firing and an answer appearing
Service topology	How it decides what else is affected, and why it will not guess
SRE doctrine	The principles it is built on, and which are not built yet
Glossary	Every term this guide uses

The shortest version

flowchart TD
    Signal["A signal arrives<br/>always as a conversation"] --> Worthy{"Worth<br/>investigating?"}
    Worthy -->|No| Log["Recorded. No model call"]
    Worthy -->|Yes| Scope["Scope: what is affected"]
    Scope --> Change["Change: what shipped recently"]
    Change --> Evidence["Evidence: logs, code, metrics,<br/>infrastructure, past incidents"]
    Evidence --> Rank["Ranked causes, each with the<br/>evidence for and against"]
    Rank --> Act["The safest next step,<br/>for a human to take"]

Scope before cause, cause before recommendation. That order is fixed, and it is the same whichever model provider you configured.

The one rule everything else follows from

Every claim is tied to a lookup that actually happened. A claim the platform cannot tie to real evidence is dropped before you see it, confidence falls when evidence is thin, and an investigation that could not reach a conclusion says so rather than producing a plausible paragraph.