How to Go with AWS Lambda?
summary
Explains how Go runs on AWS Lambda with concrete internals: Lambda downloads a zipped binary, deploys a container, bootstraps runtime, then calls the binary over a local RPC port using the open-source aws-lambda-go shim. Covers cold start lifecycle, memory→CPU tradeoffs (example: ~128MB runs a prime test ~11s vs ~1024MB ~1.6s), account soft concurrency limit ~1000 (you can throttle to 300), and packaging tips (use package individually for Go, keep binaries small). Demo shows background goroutine and reuse /tmp and global vars across warm invocations because environments freeze between calls.
AI-written summary · about
chapters (10)
- 0:00Introduction and agenda
- 1:54What is AWS Lambda
- 3:22Key differences from servers
- 7:16AWS Lambda with Go
- 9:41How AWS Lambda works internally
- 18:21Cold starts and warm functions
- 25:04Handling failures and retries
- 27:09Resource allocation and cost
- 29:26When to use Go vs Python
- 33:23Live demo and Q&A
related talks
52mGo and AWS Lambda – a perfect Match
47mReal Use Cases for Using AWS Lambda for Infrastructure Tasks
54mContainerless in the cloud with AWS Lambda
30mGo Serverless!
41mGo for serverless functions
53mBest practices for building interactive applications with Lambda
58mAWS Lambda and Serverless Framework Deep dive
39mThink in Cloud, Develop in Cloud