Upstream contribution · msitarzewski/agency-agents
agency-agents · Explain the 14.4× incident-alert burn rate correctly
Problem
An availability SLO example described its 14.4× alert as predicting budget exhaustion in two hours. An on-call engineer reading that comment would get the wrong estimate for the configured 30-day window.
Engineering response
Correct the explanation to about 50 hours at a sustained 14.4× rate. The one-hour observation window corresponds to roughly 2% of a 30-day error budget. The alert expression itself is unchanged.
Evidence and scope
The pull request shows the calculation against the example and its 6× companion rule. It reports a clean scoped lint check.