Root Cause Analysis

3rd-level Analytics & Forecasting

Casting Time: 1 hour, conducted as a meeting with all relevant parties present, or as many as could be scheduled on short notice
Range/Area: Self
Components: V, S, M (an incident timeline, incomplete)
Duration: Instantaneous
Classes: Engineering Manager, Site Reliability Engineer, VP of Engineering

Following an incident, you attempt to divine its true, underlying cause, rolling a check against your organization’s Blamelessness score. On a high enough result, the spell correctly identifies a systemic or process-level cause, along with an actionable remediation. On a failure, it instead identifies a single individual or team as the cause, regardless of whether a systemic cause also exists; that creature suffers a lasting penalty to standing the caster did not intend and will later describe as unfortunate.

This spell’s accuracy is inversely proportional to how recently the organization’s leadership was involved in decisions related to the incident.

At Higher Levels. When cast using a spell slot of 4th level or higher, you may compel the meeting’s attendees to remain until an action item is agreed upon. Nothing in this spell compels anyone to complete it afterward.

Flavor: “Action items: 1) Add more monitoring. 2) Improve documentation. 3) Revisit in Q3.” — postmortem document, zero comments, last accessed the day it was written