The team has a vague complaint: 'Deploys are unreliable.' The current dashboard shows CI uptime, runner CPU, and queue depth, but none of those state whether a developer can complete a deploy. SLO recipe: user job -> SLI -> target -> window -> budget action The common trap is choosing a metric the platform already has, such as runner uptime, instead of the behavior the product team actually experiences. User job Define the job as: an engineer with an approved change deploys a standard service to production. A clear user job prevents the SLO from drifting into internal implementation health.…
Read the full lesson
Sign up free: one personalized lesson every day, matched to your role and goals.
Already have an account? Sign in