QuestionQ74

VerifiedApplying site reliability engineering practices

You are responsible for the reliability of a high-volume enterprise application. Numerous users report that an important subset of the application's functionality — a data-intensive reporting feature — is consistently failing with an HTTP 500 error. When you investigate your application's dashboards, you notice a strong correlation between the failures and a metric that represents the size of an internal queue used for generating reports. You trace the failures to a reporting backend that is experiencing high I/O wait times. You quickly resolve the issue by resizing the backend's persistent disk (PD).

You now need to create an availability Service Level Indicator (SLI) for the report generation feature. How should you define it?

Explanation

An availability SLI should reflect the user-facing outcome rather than an internal implementation detail. Google's Site Reliability Engineering guidance defines availability SLIs as the proportion (ratio) of valid requests that are served successfully — for example, the ratio of report generation requests that return a successful response versus those that fail. Metrics like internal queue size, I/O wait time, or disk throughput are useful for root-cause diagnosis of a specific incident, but they are proxy/cause-level signals tied to one particular failure mode and current system architecture. If defined as the SLI, they would fail to detect future outages caused by different underlying issues and would not directly represent whether users are actually experiencing successful report generation. Defining the SLI as the proportion of successful report generation requests ensures it accurately and durably measures the reliability experienced by users.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!