Nothing Good Ever Happens After 2am

on YouTube ↗15m

" summary

A postmortem of a deployment-induced outage in legacy code: describes the incident timeline (detection ~24 hours after deploy), the team’s outage procedure and how SLA definitions, on-call practices and lack of a revert plan contributed to the prolonged incident, and concludes with company-agnostic operational takeaways for improving incident response.

AI-written summary · about

" chapters (5)

1. 0:00Start
2. 1:30Roles and responsibilities
3. 2:15Incident Priority Definitions
4. 2:42Tools
5. 4:42Background

" related talks

• 02mOpening RemarksMichael Chang · Reliability @Scale 2024
• 26mMission Critical Mindset Building a Culture of Resilience in DevOps TeamsNiek van Raaij · DevOpsDays Amsterdam 2025
• 22mAfter the dust has settled: learning from incidentsAjuna Kyaruzi · DevOpsDays Detroit 2023
• 28mAI-Driven Incident Response: From Alert to PostmortemPhilipp Westphalen · ContainerDays Hamburg 2026
• 25mGetting Back to Sleep ASAP for the On-Call DeveloperTom Granot · Conf42 Cloud Native 2021
• 45mFailure Happens: Improving Incident Response in Large-Scale OrganizationsDamon Edwards · LISA17
• 30mLearning From Failure: Post MortemsAlex Gaynor · PyCon 2018
• 20mHow to Have an Operational IncidentCourtney Eckhardt · DevOps Days Toronto 2019