Preparing for downtime when all you want is uptime
summary
AWS adds Incident Manager to Systems Manager to automate response plans tied to CloudWatch/EventBridge alarms, run actual or manual Systems Manager runbooks, and open Slack/Chime channels via AWS Chatbot while tracking a timeline and metrics in an incident dashboard. A real OrderDrop shows a 10% orders loss when Lambda concurrent executions hit the account limit (default 1000) because Cleanup Lambdas consumed concurrency; disabling them restores traffic. Postmortems become a blameless COE analysis with action items.
AI-written summary · about
chapters (8)
- 0:00Start
- 2:43Agenda: Let's talk about incidents
- 4:54The uptime flywheel
- 7:00Incident Manager from AWS Systems Manager
- 14:18Mechanism: Operational readiness review
- 23:40Mechanism: Correction of errors (COE)
- 28:45Takeaways
- 30:42Learn DevOps with AWS Training and Certification
related talks
31mOn the Importance of Visibility
26mA proactive approach to resilience in modern apps
20mHow to Have an Operational Incident
46mWhat You'll Miss on AWS & How To Find It
14mExpect the unexpected: Building resilience with AWS
49mMonitoring serverless applications on AWS: the monitoring monster becomes your friend.
40mSh... Fail Happens: Fail-aware Events Processing on AWS
26mHow to survive a Lambda-induced data tsunami in AWS