I've been going through interviews pretty carefully this cycle, partly because I burned out badly at my last job and I'm trying to use the process as a filter, not just a hoop. Datadog was one where the behavioral component actually surprised me in a good way.
First, the basics: there are 1-2 dedicated behavioral rounds in the full loop, plus behavioral questions woven into the hiring manager call. So you need to be ready the whole time, not just for one round.
Questions I was asked or heard about from friends who went through it: Tell me about a time you disagreed with a technical decision and what you did. Describe a project where you had to work across teams with conflicting priorities. When have you had to push back on a timeline with a manager or stakeholder? Tell me about a time you made a mistake in production and what the follow-up looked like. How do you handle it when requirements change mid-project?
What I noticed: they really lean into the cross-functional collaboration angle. Multiple questions touched on working with people outside your direct team. For a monitoring company where the product touches basically every other team at a customer's company, that makes sense.
The values Datadog seems to emphasize (per public materials and what interviewers hinted at): customer obsession, being direct and transparent, moving fast without breaking trust. The production mistake question ties directly to transparency.
My honest take: the behavioral rounds felt pretty standard STAR-method stuff. Where candidates probably stumble is not having specific examples ready, or giving vague answers that don't have a real outcome. 'We figured it out' is not an outcome. Numbers, timelines, specific decisions matter.
Also, the hiring manager call is not just a formality. Mine was 35 minutes and half of it was behavioral depth. Treat it like a real round.