Go for serverless functions

on YouTube ↗41m

" summary

Compares AWS Lambda and Google Cloud Functions for Go with concrete differences: GCP enforces an init function requirement and root-entry constraints, AWS uses generic handlers; providers reuse globals differently so global state reuse appears (stale timestamps on repeated Lambda runs) as a cold-start optimization; default GCP concurrency is single-instance, AWS allows reuse to reduce startups; timeouts run up to ~10–15 minutes; observed Lambda OOM when a 128 MB instance processed a 300 MB CSV; file-descriptor limits surface as "too many open files."

AI-written summary · about

" chapters (11)

1. 0:00Serverless functions with Go
2. 2:49Serverless key features and use cases
3. 8:54AWS Lambda vs Google Cloud Functions
4. 13:08AWS Lambda code walkthrough
5. 16:15Google Cloud Functions code walkthrough
6. 20:07Global scope and cold starts
7. 25:40Memory limits and runtime errors
8. 32:48File descriptors and connections
9. 36:10Scaling dependencies and databases
10. 37:37Key takeaways and best practices
11. 40:05Q&A

" related talks

• 30mPerformance Tales of ServerlessMikhail Shilkov · Codemotion Rome 2018
• 37mPerformance Tales of ServerlessMikhail Shilkov · Codemotion Amsterdam 2018
• 20mLost in Serverless: AWS Lambda, Google Cloud Function, Azure Functions...Sébastien Moreno · BDX I/O 2018
• 50mGo Serverless with Google CloudMarek Biskup · GDG DevFest Sofia 2019
• 58mServerless Under the HoodMikhail Shilkov · DotNext 2019 Piter
• 39mHow to Go with AWS Lambda?Taras Postument · DevFest Ukraine 2018
• 52mGo and AWS Lambda – a perfect MatchAndreas Mohrhard · DevOpsCon 2017 Berlin
• 41mServerless Compute on Google Cloud: What's NewMyles Borins, Jason Polites · Google Cloud Next '18