Observing All the Serverless Things

Explore serverless observability in AWS through this 36-minute conference talk from GOTO EDA Day 2022. Discover how to effectively monitor Lambda functions, handle failures, and leverage CloudWatch for optimal application performance. Learn about best practices, automation techniques, and the open-source tool SLIC Watch for streamlined dashboard and alarm setup. Gain insights into error tracing, capacity inspection, and making sense of alerts in serverless environments. Understand the importance of observability testing, native AWS tools, and alternatives outside the AWS ecosystem. Perfect for developers and architects looking to enhance their serverless monitoring and troubleshooting skills.

Intro
Serverless HPC FTW
A typical case study
Observability in the cloud
Can we test our ovservability?
Alerts flow
Making sense of alerts
Initial hypothesis
Gathering evidence
Checking timeouts
Inspecting DynamoDB capacity
Tracing errors
HTTP 502
Lambda CloudWatch logs
Conclusions
Before & after
What we have learned so far
AWS native observability = CloudWatch
Alternatives outside AWS
CloudWatch out of the box
Getting the best out of CloudWatch
Best practices
CloudFormation for CloudWatch alarms
Can we automate this?
Intro to SLIC Watch
How SLIC Watch works
Before & after SLIC Watch
Configuration
How to get started with SLIC Watch
What if I'm not using serverless framework?
Roadmap
Wrap-up
Outro