Cloud Cost Optimization for Java Developers
summary
Java apps waste cloud spend when JVM JIT warm‑up and over‑provisioning meet autoscaling: a 10–20k‑line Spring app needs ~10–15 minutes for 95% JIT optimization, creating sustained CPU waste; CRaC checkpoint/restore cuts Spring PetClinic cold start from ~4s to ~40ms. Moving to Arm Graviton Arm64 can shave ~10% immediately and up to ~50% vs x86 in some cases. Choose between CaaS (Cloud Run/Fargate), Kubernetes, or serverless; Google shows a modular monolith can reduce latency ~15x and cost ~9x. CRaC checkpoint/restore
AI-written summary · about
chapters (16)
- 0:00Introduction
- 2:11Speaker intro
- 5:01Cloud cost challenges
- 8:08Pricing trends and compute
- 12:44Scaling models comparison
- 18:23Architecture and elasticity
- 24:18Developer habits costing
- 29:50JVM power efficiency
- 35:50JVM startup and JIT
- 40:29Short-lived apps warm-up
- 46:46JVM optimizations
- 50:14Over-provisioning trade-offs
- 53:27Cloud functions complexity
- 56:44ARM processors savings
- 58:08Managed services costs
- 1:00:13Practical tips
related talks
51mAdapting Java for the Serverless world!
58mEnhancing Java Applications for the Cloud
47mOptimize your Java application for containers and serverless
01h 02mOptimizing Containerized #Java Apps with Jib, #GraalVM, CDS, & #CRaC
26mGoing Serverless With Spring’s Support for GraalVM, Project CraC & More
52mLean Spring Boot Applications for the Cloud
01h 03mKeeping Your Java Hot.
03h 27mAsia Pacific