Optimizing AWS Lambda for Cost & Performance Best Practices for 2023

Optimizing serverless applications have changed significantly in 2022 as teams increased serverless adoption, moved towards daily releases and a new set of tools and approaches were released to the market.
‍
As we start the new year, join Sedai to get up to speed on these key considerations to incorporate into your serverless operations for 2023. We'll be covering best practices including how to:
‍
- Prioritize cost and performance objectives at the application and infrastructure level
- Eliminate 95%+ of cold starts at scale with minimal lift and cost efficiency
- Autonomously optimize memory for AWS Lambda in real-time
- Set up autonomous optimization to achieve cost and performance goals without constant manual effort

Sedai and AWS will take you through scale your serverless operations and achieve your cost and performance KPIs in 2023. We’ll show how to virtually eliminate them effectively leveraging AWS tools and Sedai’s AI/ML-powered autonomous optimization and cloud management platform.

00:00 - Introduction
02:01 - Foundational: Lambda Execution
02:06 - Foundational: Lambda Concurrency
05:20 - Foundational: Automation vs. Autonomous
07:52 - Best Practices: Set SLOs & Priorities Aligned with Business Goals
10:45 - Best Practice: Choosing Key Metrics
13:25 - Best Practice: Optimizing Cost and Performance using Memory Settings
23:15 - Best Practice: Solve Cold Starts with Dynamic Provisioned Concurrency
29:41 - Best Practice: Set Timeouts based on Production Experience
33:03 - Best Practice: Get Real-Time Intelligence on Release Performance
37:41 - Best Practice: Leverage Low Cost Processors for Compatible Workloads
39:40 - Best Practice: Use Tracing to Drill into Performance Issues
43:45 - How safe are autonomous systems for these kinds of use systems?
45:50 - How do you do tracing for distributed environments?
46:45 - Are there any best practices for scaling across regions?
48:40 - What are some KPIs that are important to track when implementing autonomous systems in an organization?
53:05 - Can we separate the Lambda handler from the core logic?
54:10 - Are you GDPR compliant?
54:59 - What do you think about hexagonal architecture within Lambda functions?
55:38 - Any context with open telemetry?