How to Do the Same Thing Over and Over Again and Yield Different Results
summary
Automates Elasticsearch index updates using GCP Cloud Functions + Cloud Scheduler to refresh time-sensitive datasets (example: Boston Celtics game results) and also implements hobby projects like a sunset-run notifier via Render cron job and a NASA NEO pull using the NASA NEO API. Demos with Azure Functions and AWS Lambda expose real tradeoffs: different cron syntax and UTC quirks, dependency packaging (needed a pandas layer on Lambda), and occasional deployment friction, but the scheduled jobs run reliably when batched and monitored—security and cost need explicit attention.
AI-written summary · about
chapters (13)
- 0:00Scheduling automation intro
- 3:13Why scheduling matters
- 5:42Cron vs serverless functions
- 8:10CRON syntax explained
- 9:56Running reminder bot
- 12:07Risks of automation
- 13:28NASA asteroid data fetch
- 15:58GCP automation setup
- 18:50Azure Functions deployment
- 22:38AWS Lambda scheduling
- 25:04Cloud scheduling comparison
- 26:51Key takeaways
- 27:32Closing
related talks
48mServerless design patterns
49mGet silly with serverless
39mHow to Get Started with Server-less on Google, Amazon & Microsoft
54mFocus on Your Code, Not Infrastructure, with Google Cloud Functions
35mConvergence of Serverless APIs & Compute
37mPerformance Tales of Serverless
23mnow with Functions!
38mChoreography vs Orchestration in serverless microservices