Technical depth is easy to test for. The ability to stay useful at 3am during a Sev1 is not - but there are interview signals that predict it.

Most interview loops for incident and problem management roles test the wrong thing. They ask candidates to describe an outage they handled, and the candidate performs competence: the tools they used, the RCA they wrote, the MTTR they improved.

That tells you almost nothing about how they behave when the bridge has 40 people on it and nobody knows what broke.
Three signals that have actually predicted it for me:
They narrate uncertainty without apologising for it. Ask what they did when they did not know the cause yet. Strong candidates describe saying so out loud and stating what they were doing next. Weak ones skip straight to the resolution, because in their retelling they were never lost.
They separate restoration from root cause without being prompted. Anyone who has run a real Sev1 knows these are different jobs with different clocks. Candidates who conflate them have usually only written the post-incident report, not run the call.
They can name a decision they got wrong on a live call. Not a tooling gap, not a process gap. A call they made that was wrong, and what it cost. Calm under pressure is not the absence of mistakes, it is the ability to keep working after making one in front of an audience.
None of this is testable from a resume, and all of it survives a 20 minute conversation.