Interview scorecard
Site Reliability Engineer
Score each skill from 1 to 5. A 1 is below the bar for this role, a 3 meets it, and a 5 is above it. The examples show what each sounds like. Use 2 and 4 for answers that fall in between.
Reliability engineering
Designs and improves services so they meet agreed uptime targets and fail in safe, predictable ways.
Questions
- Tell me about a time you made a service more reliable. What did you change?
- How have you set or used a reliability target, like an SLO, for a service?
- Describe a failure you expected in a system. How did you design for it?
1: Below the bar
Describes fixing outages one at a time. Has no reliability targets and can't say what made the service more reliable.
3: Meets the bar
Names a clear improvement and a reliability target. Explains some causes of failure but has limited proof the change worked.
5: Above the bar
Ties changes to a target and measured results. Plans for expected failures, such as retries or fallbacks, and tests them before they hurt users.
Incident response
Restores service quickly during outages, communicates clearly, and follows up so the same failure doesn't recur.
Questions
- Tell me about a serious outage you handled. What did you do first?
- Tell me about a time you updated stakeholders mid-incident. What did you say?
- Tell me about a post-incident review you ran. What changed afterward?
1: Below the bar
Dives into fixes without a plan or updates. Blames people afterward and can't name any lasting change.
3: Meets the bar
Restores service by following the runbook and keeps others informed. Writes a review with some follow-up actions.
5: Above the bar
Stops the damage first, assigns roles, and updates stakeholders on a schedule. Drives follow-ups that prevent repeats.
Monitoring and alerting
Designs signals and alerts that catch real user-facing problems early without waking people for noise.
Questions
- Tell me about a time you set up monitoring for a service. What did you choose to measure?
- Tell me about an alert that fired too often or too late. What did you change?
- Describe an outage your monitoring missed. How did you find out, and what did you add afterward?
1: Below the bar
Monitors only basic machine stats like CPU. Treats noisy alerts as normal and can't say how an alert helps users.
3: Meets the bar
Picks sensible signals tied to user impact. Tunes noisy alerts after complaints and adds a missing check after an outage.
5: Above the bar
Starts from user impact and sets alert thresholds with data. Regularly prunes noisy alerts and closes gaps before they cause outages.
Automation
Replaces repetitive manual operations work with reliable, maintainable scripts and tools that others can safely use.
Questions
- Tell me about a manual task you automated. How did you decide it was worth it?
- Tell me about automation you built that failed or caused a problem. What happened?
- How did you make sure your team could run and maintain what you built?
1: Below the bar
Automates only when told to. Describes one-off scripts that only they can run, with no testing or documentation.
3: Meets the bar
Picks repeated tasks to automate and builds working tools. Adds some testing and notes, but safety checks are limited.
5: Above the bar
Weighs time saved against upkeep, builds in safeguards and rollback, and gets teammates using and improving the tool.
Capacity planning
Forecasts resource demand and acts early so services stay fast and affordable as usage grows.
Questions
- Tell me about a time you predicted a service would run out of resources. What data did you use?
- Describe how you decided how much extra capacity to keep as a safety margin.
- Tell me about a forecast that turned out wrong. What did you change afterward?
1: Below the bar
Reacts after resources run out. Relies on guesses and can't name the data behind any capacity decision.
3: Meets the bar
Uses past usage trends to forecast and add capacity in time. Sets a margin but can't fully justify it.
5: Above the bar
Combines usage trends with upcoming launches, tests limits under load, justifies margins by cost and risk, and corrects forecasts.
Overall recommendation
Decide on your own, before you compare notes with the rest of the panel.