Google Site Reliability Engineer Interview Questions

I would describe the incident, what page or alert triggered it, and the impact on users. Then I would explain how I diagnosed the problem, what actions I took to reduce the blast radius, and how I communicated with the team. The reason I choose this structure is that SRE interviews care about calm incident handling as much as technical debugging. I would also mention what I learned and what I changed afterward to prevent the same issue from recurring. If the interviewer wants more detail, I would talk about escalation, timelines, and postmortem follow-up.

An SLI is the metric that tells us how the service is performing, such as latency or availability. An SLO is the target we want that metric to meet, while an SLA is the formal promise we make to users or customers. Error budget is the allowed gap between perfect reliability and the SLO target, and it helps balance reliability work with feature work. The reason I would explain them this way is that the terms only make sense when they are tied to operational decisions. If the interviewer wants more detail, I would give an example from a real service and show how the error budget changes release choices.

I would start by checking whether the slowdown is in the application, the database, the network, or the infrastructure layer. Then I would look at recent deploys, traffic changes, saturation metrics, and traces to find the hottest path. The reason I choose this approach is that latency problems are usually caused by one or two bottlenecks rather than the whole stack. I would also think about whether the issue is limited to one region, one endpoint, or one class of requests. If the interviewer wants more detail, I would explain how I would isolate the problem step by step.

I would walk through DNS lookup, TCP or TLS connection setup, HTTP request delivery, server response, and browser rendering. The reason I choose this answer is that it tests whether I understand the full request path, not just one layer. I would also mention caching, redirects, and how the browser decides whether to reuse an existing connection. If the interviewer wants more detail, I would explain the difference between a cold load and a warm cached load. The key is to show that I can reason from the client all the way to the server.

I would define the most important SLIs first and build dashboards and alerts around those metrics. Alerts should focus on user impact and be actionable, not just noisy. The reason I choose this design is that a good monitoring system helps the team notice real problems quickly without creating alert fatigue. I would also think about severity levels, routing, and how to avoid duplicate or low-value alerts. If the interviewer wants more detail, I would discuss how I would validate that the alerts actually catch meaningful incidents.