CloudWatch-ing: Creating More Useful Logs & Alerts with AWS
summary
AWS CloudWatch replaces scattered per-machine .txt logs with centralized CloudWatch log groups so all instances and threads are searchable, cutting batch lag from ~30 minutes to about five. Uses structured log levels (debug/info/warn/error), Terraform-created incremented metric filters that turn specific messages into graphable metrics, and SNS→Slack alerts for errors. Handles a FedEx week-long outage down to a 1–2 hour incident, and avoids CloudWatch’s 90,000 msgs/min throttle by aggregating counts before sending.
AI-written summary · about
chapters (7)
- 0:00Start
- 1:27WHY SHOULD YOU LISTEN TO ME?
- 1:42VISIBILITY WAS MY 1ST REAL PROJECT
- 2:07STORY TIME
- 8:14ENTER AWS CLOUDWATCH!
- 13:58CENTRALIZED LOGGING
- 39:48LOG STRATEGICALLY
related talks
49mMonitoring serverless applications on AWS: the monitoring monster becomes your friend.
28mPatterns for scalable security detections on AWS
30mBest practices for application observability at scale
50mTo Normalized Logs, and Beyond - Building a Threat Detection...
28mPatterns for scalable security detections on AWS
24mObservability with AWS
51mUsing observability to build trust & improve incident response times
20mDesigning the Right Observability Stack in AWS Cloud