A production service started showing errors and timeouts after a deployment. How would you troubleshoot it?

"I interviewed for a Google Site Reliability Engineer role at the Software Engineer III level in EMEA, and the process was more specialized than a normal SWE loop. I first had a recruiter call, then a 45 minute coding and scripting screen, and after that I went through troubleshooting, googliness, and Linux-heavy technical rounds. The biggest thing I noticed was that even the simpler questions kept getting reframed around production scale, Linux depth, and whether I understood why something works, not just how to use a tool or command. The troubleshooting round felt the most realistic because it was run like a simulated incident, and the Linux questions went very deep across the stack. One thing I think people should know is that even if you clear the interviews, team matching can still take a very long time depending on headcount and fit."