Boost your serverless application availability with AIOps
summary
AWS pitches Amazon DevOps Guru as a fully managed AIOps layer that ingests CloudWatch metrics and logs, X-Ray traces and CloudFormation events to detect, correlate and explain serverless failures. Demo shows a deliberate change—DynamoDB read capacity 5→1—that triggers throttling, rising Lambda invocations and API Gateway 5XXs; DevOps Guru surfaces an insight with top three metrics, related CloudFormation event, log samples and a rollback recommendation. Enrollment is a few clicks, baselining takes 24–48 hours, and log anomaly detection currently covers Lambda and DynamoDB.
AI-written summary · about
chapters (9)
- 0:00Introduction and agenda
- 2:14IT operational challenges
- 9:29Introducing DevOps Guru
- 14:04How DevOps Guru works
- 16:31Log anomaly detection
- 18:28Serverless use case demo
- 26:25Proactive insights
- 27:45Enabling DevOps Guru
- 32:45Q&A and closing
related talks
39mObservability in DevOps
35mBring the power of machine learning to your developer operations
40m4 Key takeaways from running Serverless on production for 4 years
35mBring the power of machine learning to your developer operations
26mA proactive approach to resilience in modern apps
29mHow is the role of DevOps changing in a Serverless environment?
44mApplying principles of chaos engineering to serverless
51mTaking Serverless to the Next Level