You're on call and an alert fires for an outage in a service you've barely touched. Your senior is offline. What do you do in the first 15 minutes?
situational · Junior level · general
What the interviewer is really asking
Assess whether a junior can stay calm in an incident, follow the runbook, and escalate quickly rather than freeze or hero-fix something they don't understand.
What to say
- Stabilize first: pull up the runbook and dashboards, confirm the alert is real and gauge user impact before changing anything.
- Escalate early and widen the call — page the secondary on-call or open an incident channel rather than soloing a service you don't know.
- Keep a running timeline and post status updates so whoever joins can catch up in seconds.
What to avoid
- Don't say you'd start changing production config or restarting things to 'see what happens' before you understand the blast radius.
- Don't say you'd wait quietly for your senior to come back online while users stay impacted.
- Don't downplay it — never assume the alert is a false alarm without checking the dashboards first.
Example answers
Strong: I'd acknowledge the page so it stops escalating, then open the runbook for that service and the dashboards to confirm it's a real outage and roughly how many users are hit. Since I don't know the service well and my senior's offline, I'd page the secondary on-call within a couple of minutes and open an incident channel, dropping in what I see: which alert, the error rate, when it started. I'd avoid touching prod until someone who knows the system is on, and I'd keep posting updates so they can catch up fast.
Weak: I'd start restarting the service and rolling back recent deploys to try to fix it before anyone noticed, then tell my senior once it was back.