Incident response: a product tour
What does this engineer do when production breaks, and do they trust the AI assistant’s suggestion?
You’re Alex Moreno
Candidate
Alex Moreno is on call when checkout starts failing. Then you see what the hiring team sees: the level for each skill, the actions behind it, and what the rules used.
9 steps · about 2 minutes
Read the tour as text
What do they check first?
Alex is on call and checkout is failing. The first recorded step is Check this metric on checkout success, which shows about 18% of checkouts failing.
Do they find where it breaks?
Alex opens a healthy checkout request and a failing one and compares them. The failing one breaks inside Checkout before any payment is attempted.
Do they trust the AI assistant?
The AI assistant in the channel suggests scaling Checkout. Alex rejects it with a one-sentence reason from the evidence. Approving a wrong suggestion is recorded too.
Do they confirm before changing production?
Alex reads the runbook, finds the flag that went to every shopper at 14:10, and runs a read-only command to confirm it before touching production.
Do they verify the fix?
Alex tells the channel first, turns the flag off, waits for the change to land and checks the same customer metric again. Then Alex posts that checkout has recovered.
What does the candidate see at the end?
In hiring, Alex gets a submission receipt, not levels or feedback, so the answers cannot be passed on.
How is each level set?
In an incident assessment you publish now, rules set every level from what Alex did, and no model reads the work. Each level says why and lists the events it cites.
What did the rules use?
You see what the rules used: the time to mitigation, production changes that did not help, wrong AI suggestions that were approved, and how much of what Alex opened was relevant. Opening everything does not count as choosing what to look at.
Who decides?
A person. The timeline shows every action in plain words, and the actions a level cites are highlighted. You read it and decide.