The study analyzed 366 incident resolution attempts in a Linux training environment between October 10 and November 8, 2026. Out of 775 initial requests, 752 environments were successfully prepared, but only 366 users proceeded to start an incident.
The data shows that 246 attempts reached the verification stage, with 192 resulting in successful service restoration, representing a 52.5% success rate for all attempts and a 78% success rate for those that reached verification.
Performance varied by scenario: Kubernetes CrashLoopBackOff had a 90% success rate, while general application failures were resolved in 42.6% of cases. OOM Killer and Too many open files scenarios showed success rates of 40% and 42.9% respectively.
The 502 Bad Gateway scenario was resolved in 12 out of 14 attempts, with a median resolution time of 13 minutes and 57 seconds. The system recorded 174 attempts that ended without successful service restoration.
