What defines toil, with an example?

WHAT THIS TESTS The interviewer wants to know whether you can recognize the specific kind of operational work that SRE exists to eliminate, and whether you understand why bounding it protects time for engineering. It also probes whether you can be concrete rather than abstract.

A GOOD ANSWER COVERS Define toil precisely. It is work tied to running a service that is manual, repetitive, automatable, tactical or interrupt-driven, has no enduring value, and grows at least linearly as the service scales. Emphasize that toil is not the same as overhead such as meetings, email, or hiring, which are necessary but unrelated to running the service. Mention that Google's guidance is to keep toil below roughly half an SRE's time so the rest goes to engineering that reduces future toil. Then ground it in a concrete example.

COMMON WRONG ANSWERS Calling all operations work toil; toil is a specific subset. Equating toil with anything boring or unpleasant. Confusing toil with overhead. Claiming hard, novel project work is toil simply because it is operational.

LIKELY FOLLOW-UPS How would you measure toil across a team? How do you decide which toil to automate first? When is automating something not worth the cost? What is the danger of zero toil?

ONE CONCRETE EXAMPLE A service occasionally hangs and the runbook says SSH in and restart it, happening a few times a week and waking on-call. It is manual, repetitive, automatable, has no lasting value, and worsens as the fleet grows. The fix is a health check plus an automatic restart, and ultimately fixing the root-cause leak. That progression, recognize the toil, automate the response, then eliminate the cause, is exactly the judgment the question rewards.