A hitchhicker's guide to Serverless computing
summary
Demo app 'pic-a-daily' is a photo‑sharing PWA: uploads land in Cloud Storage, trigger Cloud Functions (Node) that call Vision API, update Datastore, and Pub/Sub fires a periodic tag‑cloud function; front end runs on App Engine (Java) and admin on Go. Stackdriver provides logs, traces and an SLO dashboard that surfaces a recurring ~3s latency spike from cold starts; App Engine versioning lets you split traffic and migrate/rollback instantly. Argues serverless (including Knative containers) supports scale to zero and needs SRE practices like SLIs, SLOs and an error budget.
AI-written summary · about
chapters (14)
- 0:00Start
- 4:51Google Cloud
- 8:53Containers
- 13:08Auto Scaling
- 17:17Event-Driven Logic
- 18:35How Do You Leverage Google's Experience of Running Services at Massive Scale
- 23:02Infrastructure
- 26:11Service Level Indicators
- 29:25Versioning
- 31:11Tools
- 33:16Admin Tool
- 41:00Recap
- 47:19Vision Analysis
- 49:13Is It Possible To Have an Sre That Looks after Two Products or Two Different Teams
related talks
43mGoogle Cloud's new serverless features
30mThe Hitchhiker's Guide to the Serverless Galaxy
43mWhat's new in Google Cloud Serverless
39mBuild a modern Java Serverless Application
42mNew serverless solutions on Google Cloud for functions, apps and containers
02h 47mBuilding a Photo Sharing Service with Google Cloud Serverless Tech
40mWhat's new in Google Cloud serverless
35mIt’s Easy to Say Serverless