利用人工智能驱动的可观测性增强 DevOps 能力.pdf

《利用人工智能驱动的可观测性增强 DevOps 能力.pdf》由会员分享，可在线阅读，更多相关《利用人工智能驱动的可观测性增强 DevOps 能力.pdf（73页珍藏版）》请在三个皮匠报告上搜索。

1、 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.DEV304Supercharge DevOps with AI-driven observabilityRossana Surez(she/her)Tech Lead NaranjaXAWS C

2、ontainer HeroElizabeth Fuentes Leone(she/her)Developer Advocate/SDE,GenAIAmazon Web Services 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.AgendaThe limits of traditional observabilityAI-powered observability in CI/CD(GitHub Actions)AI Agent+Demo in actionThe Impact,Lessons&Futu

3、re 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.“everything fails,all the time.”Dr.Werner VogelsVP and CTO at A 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.2025,Amazon Web Services,Inc.or

4、its affiliates.All rights reserved.2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.The limits of traditional observability 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.Why traditional observability falls shortTraditional observability=reactive firefighting,no

5、t proactive prevention.Reactive,not proactive 2 min detection after user complaint1Noise overload 200 alerts in 5 min,99%noise2Siloed signals 3 tools,0 correlation3Slow decision-making 40 min debating,10k failed4 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.Impact on teams&busi

6、nesslost/hour70%Time firefightingAlert fatigue40%$50K$500K Revenue impactBurnoutLost innovation 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.The gap we need to bridgeLogs,metrics,traces everywhere.200 alerts,3 tools,no correlationSlack war room 40 min to decide rollbackCustomer