Lightning Talk: Searching for Serverless Candidates in the Jungle of Microservices
Lightningat AppDeveloperCon NA 2024 · on Tuesday, 2024-11-12· at 17:15· Salt Palace | Level 1 | 151 G·
summary
Pipeline collects resource metrics via Prometheus, builds an interaction graph from network traces, decomposes each time series with seasonal ARA, and runs spike modeling to mark active versus idle intervals. Applied to 50 services across 10 metrics, the process flags 19 bursty/seasonal services and narrows real re‑architecture candidates to 5. Conversion options include Kubernetes Jobs, split control‑plane + batch workers, or a hybrid cron+replica setup. Results show $3,200 saved across three apps, ~20% COGS improvement and 15% fewer container restarts, plus real‑time anomaly alerts.
AI-written summary · about
chapters (9)
- 0:00Serverless and microservices intro
- 1:03Serverless vs microservices
- 2:26Start with microservices later optimize
- 3:51Finding serverless candidates
- 4:33Data-driven approach
- 6:46Analyzing resource spikes
- 7:46Case study results
- 9:57Key outcomes improvements
- 10:19Q&A
related talks
34mFleeting Metrics: Monitoring Short-lived or Serverless Jobs with Prometheus
11mLightning Talk: Grasp the Real Worth of Your Microservices by Brewing a Truth Poti...
01h 02mServerless Architecture Patterns
46mServerless Deployment of Analytics
36mTransforming SaaS with Serverless Architectures
47mMonitoring in the Microservices world: Kubernetes vs Serverless based applications
31mLightning talks - backend
23mLightning Talk: Rethinking Microservices