[Claude code 개발자 Boris Cherny, 소스코드 유출 경위 공개
Claude Code 개발자 Boris Cherny가 3월 31일 발생한 Claude 서비스 장애(Opus 4.6, Sonnet 4.6 타임아웃 급증)에 대해 원인을 공개했다. 장애는 **수동 배포 단계의 미흡한 자동화**가 원인으로, Cherny는 개인의 실수가 아닌 **프로세스/문화/인프라의 문제**로 정의하며, **블레임리스 포스트모템(blameless postmortem)** 철학을 적용했다. 장애는 12시간 이상 지속되다 완전 복구됐으며, 팀은 자동화 개선을 진행 중이다. 이 사건은 AI 서비스에서 **신속한 장애 대응과 시스템 개선**의 중요성을 강조한다.
Игровые серверы на Cozystack: первоапрельская нешутка
Привет, Хабр! Мы — команда Cozystack , open-source платформы для построения облаков на своём железе. Хотим рассказать, почему мы решили целиться в направление игровых серверов и что из этого вышло.
https://habr.com/ru/companies/aenix/articles/1018034/
#cozystack #aenix #игровые_серверы #devops #kubernetes #platform_engineering #cloud #облачные_технологии #геймдев #sre
WikiApiary is back online after having been down and overloaded for a long time.
Bawolff brought it back and did a ton of full stack performance work to optimise the whole system. Adding HTTP caching, tuning InnoDB, and more!
https://blog.bawolff.net/2026/03/giving-wikiapiary-kick.html
#MediaWiki #WikiApiary #Wikibase #Wikidata #VinylCache #VarnishCache #MariaDB #MySQL #webperf #SRE
System Administration: Week 9: Writing System Tools
This week, we're discussing a few core concepts in software engineering, the evolution of your typical sysadmin's tool chest from shell aliases and one-off scripts to system utilities shared with the org to full fledged software components. We also cover good coding practices, bug reporting, commit messages, ... and, ugh, how all that is touched by AI now and what that means for us.
✍️ Mon CV a un SLA de 99.9% de fraîcheur.
Chaque nuit, une GitHub Action met à jour mes stats, rebuild le PDF (LaTeX), génère une version HTML et publie le tout.
Un CV toujours à jour, versionné, automatisé.
"Absolutely not." That was the starting position on AI.
Bastian Spanneberg's Ignite at DevOpsDays Zürich 2026: a reluctant convert's journey from AI skeptic to finding LLMs surprisingly effective for SRE work.
Dislike doesn't change reality. So the internal blockade had to go.
https://www.devopsdays.ch/event/program/ignites/bastian-spanneberg/
Struggling with Kubernetes Pod failures in production?
Learn how to troubleshoot CrashLoopBackOff, Pending Pods, ImagePullBackOff, and OOMKilled errors using real-world DevOps debugging techniques and kubectl workflows.
#Kubernetes #DevOps #CloudComputing #SRE #K8s #Observability #PlatformEngineering #tech
Zero downtime isn’t a feature — it’s an architecture.
Explore how AWS enables resilient banking systems in 2025 using Aurora failover, ECS orchestration, Blue-Green deployments, and canary releases to eliminate service disruptions and protect transactions.
sayzard
Habr
Timo Tijhof
𝚓𝚓@localhost$:~
Jan Schaumann
deliverator
Bruno Adelé
DevOpsDays Zurich
Ismail Kovvuru
