Mode Select
Choose Difficulty
Scenario List
SRE Incident Environment
Select a difficulty tier, pick a scenario family, inspect alerts, rollout state, and trace clues, then compare your response against the seeded benchmark.
Selected Scenario
Choose a scenario from the selected tier to begin.
Investigation now matters: versions and dependencies start hidden, metrics must be inspected directly, and hard mode requires an explicit recovery ping before final diagnosis.
Live State
Operator Controls
Suggested flow: inspect metrics or logs, discover dependencies, apply a fix only after evidence, then validate recovery before submitting diagnosis.
Queue
AI Runtime
Configure provider credentials and model settings in the browser, run AI baselines or full benchmarks, and compare the configured agent against the current human session without changing the parent console flow.
Config
Actions
The AI runtime page sends provider configuration per request so the current console and existing scripted/human flows remain unchanged.
Results
Results
Comparison
Activity